Computer Science
See recent articles
Showing new listings for Monday, 16 November 2026
- [1] arXiv:2611.17338 [pdf, html, other]
-
Title: When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver ErrorsComments: 12 pages, 2 figuresSubjects: Artificial Intelligence (cs.AI)
LLM memory systems often delete facts classified as stale. We show that this hard-filtering step can reduce downstream accuracy when resolver errors delete query-critical facts. Across three models, a controlled sensitivity intervention moves hard-drop from gains of +0.186 to +0.230 at perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability. Paired targeted interventions show that false-drop causes significantly greater degradation than false-keep, while a non-deleting baseline can still fail when incorrect labels steer generation. On a frozen 102/102 calibration/test split, a simple deployment-matched gate reaches 0.655 accuracy versus 0.641 for always hard-drop and 0.662 for test-informed oracle selection. Global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. Deletion should instead be treated as cost-sensitive: validate query-critical false-drop risk on matched data, or default to label-free retention when such validation is unavailable.
- [2] arXiv:2611.17339 [pdf, other]
-
Title: Does the Ranking Hold? A Practitioner-Facing Audit of Hallucination Detectors on Small, Quantized, On-Device LLMsSubjects: Artificial Intelligence (cs.AI)
Hallucination detection for large language models is dominated by methods evaluated on 7B-70B research models with abundant sampling budgets and exact-match or ROUGE correctness labels. The small models practitioners actually ship on-device, Phi-3-mini, Llama-3.2-3B-Instruct, Qwen2.5-3B-Instruct, and Gemma-2-2B-Instruct, typically at 4-bit precision, sit outside this evidence base. We present a reproducible one-GPU audit of four detectors spanning three paradigms: length-normalized token-margin calibration, a lexical n-gram baseline, and two sampling-based consistency methods (semantic entropy and SelfCheckGPT-NLI) on two small open instruction-tuned models (Llama-3.2-3B-Instruct and Qwen2.5-3B-Instruct) loaded at 4-bit NF4 quantization on a single Colab T4 GPU, evaluated on closed-book TriviaQA. Cross-model ranking transfer is poor: Spearman rho is -0.80 and Kendall tau-b is -0.667, meaning the detector ranking actively inverts between the two models, and the cheap lexical baseline is statistically indistinguishable from the sampling-based consistency detectors on both models under bootstrap 95 percent confidence intervals. As a second, methodological contribution, we show that the standard token-overlap-to-multiple-choice correctness labeler for TruthfulQA-Generation produces near-chance AUROC (0.46-0.59) for all four detectors on both models, a labeler failure, not a detector failure, since the same detectors reach 0.69-0.89 AUROC on TriviaQA under the identical pipeline. We release the harness, prompts, and per-item scores and frame the detector choice explicitly as a fairness-of-error-rates allocation problem for practitioners deploying on resource-constrained, on-device hardware.
- [3] arXiv:2611.17340 [pdf, html, other]
-
Title: LoRAWeaver: Directional Adapter Routing in LoRA Representation SpaceSubjects: Artificial Intelligence (cs.AI)
Large language models are expensive to update through full retraining. Parameter-efficient fine-tuning reduces the cost of adaptation, but repeated adaptation creates a growing pool of task-specific LoRA adapters. This pool complicates adapter selection during inference. Existing routing and encoder-based methods often require retraining when new adapters are added, which limits their use in dynamic settings. This paper treats adapter routing as a representation alignment problem in the adapter-output space. The proposed training-free method selects adapters by matching adapter-conditioned query representations against task prototypes using cosine similarity. Experiments with LLaMA-2-7B and LLaMA-2-13B on a 48-task FLAN-v2 benchmark and on out-of-distribution tasks show that the method outperforms training-free baselines in the cross-task setting and approaches a trained retrieval encoder without learning a router. An ablation confirms the benefit of directional alignment in the adapter-output space.
- [4] arXiv:2611.17341 [pdf, html, other]
-
Title: Tabular Population Priors as Structured Retrieval for EHR Medication Set PredictionComments: Accepted at the ICML 2026 Workshop on Structured Data for Health (SD4H), Seoul, South KoreaSubjects: Artificial Intelligence (cs.AI)
Medication prediction from electronic health records is a fixed vocabulary structured prediction problem, yet recent clinical RAG systems often treat it like open ended case based reasoning: retrieve similar patient admissions and ask an LLM to infer a medication set. This retrieval object is indirect the model receives retrieved patient records while the target is a discrete set of medication class labels. We propose a more task aligned retrieval object: a tabular population prior ranked medication class candidates from a sparse condition medication co occurrence table estimated from training admissions. Using PACERAG as a fixed downstream pipeline and replacing only its patient retriever, this drop in swap improves F1 from 0.303 to 0.338 (+11.5% relative) and Jaccard from 0.218 to 0.237 on 1,004 heldout MIMIC IV admissions, with no additional LLM calls and no inference time patient index. A non LLM top k baseline reaches F1=0.248, confirming the prior is predictive before LLM integration. Aggregate population statistics can provide a more task aligned, efficient, and auditable retrieval source than patient level nearest neighbors for fixed vocabulary EHR set prediction.
- [5] arXiv:2611.17355 [pdf, html, other]
-
Title: Everything is an Operational View: UI as a Projection over Context in Agentic SystemsComments: 9 pages, 2 figures, ongoing workSubjects: Human-Computer Interaction (cs.HC); Software Engineering (cs.SE)
The emergence of generative AI has enabled Generative User Interfaces (GenUI) that dynamically construct interfaces from user intent, yet existing approaches rely on request-response pipelines that fragment context and limit transparency and control. This paper proposes AFS-UI, which reconceptualises UI as projections over context built on an agentic file-system abstraction that organizes data, capabilities, and interaction state within a unified, persistent workspace. In particular, AFS-UI addresses key limitations in current systems, including heterogeneous resource discoverability, incompatible operation semantics, tangling projection and context, rigid source composition and substitution, and limited evolvability. Within this framework, interfaces become dynamic, inspectable views over shared resources, supporting both human and agent interaction through a common operational surface.
- [6] arXiv:2611.17359 [pdf, html, other]
-
Title: fieldsmith: Automated symbolic derivation in classical field theory and gauge theory from LaTeX inputComments: 40 pagesSubjects: Symbolic Computation (cs.SC); High Energy Physics - Theory (hep-th)
We present fieldsmith, an open-source Python library for the automated derivation of analytical results in classical field theory, and non-abelian gauge theories. A Lagrangian is entered the way it is written on a blackboard, in LaTeX index notation, and the library produces and verifies Euler-Lagrange equations (of arbitrary derivative order, on flat or curved backgrounds), Noether currents and identities, canonical and Hilbert energy-momentum tensors, canonical momenta and Hamiltonians, local Poisson brackets and a Dirac-Bergmann constraint analysis with degree-of-freedom counting, linearised wave equations and dispersion relations, gauge-covariant derivatives and field strengths for products of unitary groups, and mass spectra after spontaneous symmetry breaking. A fitting routine translates component-level output back into covariant index notation, and a grading module checks students' submitted field equations, identities and conserved currents. The implementation is a compact (about 2,700 lines) layer on top of SymPy: tensors are stored by components with labelled, variance-aware slots, and every statement is established by exact symbolic computation rather than pattern matching. This document is the companion guide to the library. It explains the intent and design, describes the algorithms, and then works through eighteen self-checking examples, ranging from the Klein-Gordon field to the BPST instanton and the 't Hooft-Polyakov monopole, each preceded by the background physics needed to follow it. The examples double as a compact introduction to classical field theory and gauge theory as seen through a computational lens. We are explicit about the limitations of the approach.
- [7] arXiv:2611.17377 [pdf, html, other]
-
Title: Beyond Attack-Based Accuracy: Rethinking Adversarial Robustness Evaluation Through Structural Gradient AlignmentSubjects: Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV)
Visual adversarial robustness remains a fundamental security property as vision systems evolve from conventional classifiers to vision-language models (VLMs) and large vision-language models (LVLMs), with visual inputs remaining an attack surface across generations. Yet existing evaluation methods often depend on prescribed attacks, in-distribution data, extensive probing, or classifier-specific formulations, limiting external auditing of diverse models. We introduce Structural Gradient Alignment (SGA), a complementary robustness metric measuring the structural alignment between input content and objective-conditioned local model sensitivity. SGA uses a few randomly sampled, unlabeled public probes that need not follow the downstream task distribution and only one input-gradient computation per model--probe pair. Across 38 distinct DNNs, VLMs, and LVLMs, SGA produces rankings consistent with attack-based references. Without robustness references, uncalibrated SGA achieves $83.72\%$ average pairwise agreement across eight public-probe conditions, outperforming the strongest overall baseline in all eight by $9.38$ percentage points on average. With known robustness orders of anchor models, calibrated SGA increases agreement to $85.36\%$. These results demonstrate an attack-agnostic, computationally efficient robustness signal across conventional DNNs, VLMs, and autoregressive LVLMs, downstream tasks, threat norms, probe distributions, and evaluator-knowledge settings. Our implementation is available at this https URL.
- [8] arXiv:2611.17381 [pdf, other]
-
Title: ProdPilot: A Layered Model Context Protocol System for Automated Production Readiness and Deployment of Vibe-Coded ProjectsSubjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI)
AI-assisted development environments turn natural-language requests into working applications, yet the projects they produce often fail on their first deployment for operational reasons: credentials in source files, missing security middleware, no container definition, no health endpoint and no delivery pipeline. Automated repair of such defects usually relies on free-form model generation, which varies between runs and is seldom checked independently. This paper presents ProdPilot, a local Model Context Protocol server that audits this http URL and React projects against 50 production readiness rules, repairs violations through a bounded loop, gates the result and deploys it. Each rule is classified by what its fix requires: 28 fixes are fixed templates, 16 are filled with values extracted from the project and only 6 are written by the IDE agent's model, so 88.0\% of the ruleset needs no model-written content, and across 499 real repositories 93.5\% of the violations found fall in those classes. A fix, however it was produced, is accepted only if the checker that flagged the rule now passes. In a full-chain run over 23 sample projects, template and extraction fixes that were applied as written passed their checkers in 87.6\% and 95.0\% of cases, and independent verification rejected 25 of the 237 fixes reported as applied. A deployability model trained on 684 real deployments ranks held-out projects at 0.909 ROC AUC, a gated project was deployed, smoke-tested and connected to an active CI/CD pipeline in 82 seconds, and the unchanged server worked from the IDE agents in VS Code, Cursor and Windsurf, where live agents resolved all six delegated rules on the first attempt under the same verification. ProdPilot itself runs no model.
- [9] arXiv:2611.17386 [pdf, html, other]
-
Title: TSP instances that require exponentially many subtour elimination constraintsSubjects: Discrete Mathematics (cs.DM); Combinatorics (math.CO)
We construct pathological TSP instances that require exponentially many subtour elimination constraints (SECs) in the sense that, if a smaller collection of SECs is imposed, then the Dantzig-Fulkerson-Johnson model returns a disconnected solution costing less than an optimal tour. For asymmetric instances on $n$ vertices, we prove a growth rate of \(\Omega(\phi^n)\), where \(\phi\approx 1.618\) is the golden ratio. A node-splitting reduction gives metric instances with \(\Omega(\phi^{n/2})\) growth. We also give instances in the plane under the Euclidean, Rectilinear, and Maximum metrics that require exponentially many SECs. Concorde struggles on them, with solve times growing exponentially despite tiny integrality gaps.
- [10] arXiv:2611.17388 [pdf, html, other]
-
Title: Time-Consistent Evaluation of Hierarchical Code Graphs for SWE AgentsXianpeng (Simon)Sun, Tian Yu, Haonan Sun, Sheng Ma, Qincheng Zhang, Lifei Rao, Jie Li, Jian (Ken)Cao, Qi ZhangComments: 10 pages, 5 figures, 6 tablesSubjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI)
Evaluating repository-aware software engineering (SWE) agents requires strict separation between pre-existing repository knowledge and future code evolution, yet standard benchmarks frequently suffer from temporal contamination and prompt-level solution leakage. Building on the time-consistent benchmark protocol of Sun (2026), we enforce strict zero-leakage isolation across DragonFly in C++ and React in JavaScript/TypeScript, locking repository snapshots and knowledge graphs at historical cutoff commits $T_0$ and evaluating 100 post-$T_0$ pull requests across four prompt granularities: Minimal, Concise, Contextual, and Guided.
Under this zero-leakage protocol, we conduct a matched evaluation of three context-augmentation regimes: (1) an unaugmented Baseline, (2) a flat syntax-level Tree-sitter AST graph, and (3) ServiceExplorer, a deterministic two-level hierarchical dependency graph capturing Level-1 system/package containers and Level-2 architectural modules. Across 7,200 paired evaluations spanning six foundation models across the MAI-Code Flash, Claude, and GPT-Sol families, ServiceExplorer consistently outperforms both Baseline and Tree-sitter AST across every model, repository, and prompt granularity ($+15.6\%$ to $+86.2\%$ relative $F_1$ lift, $p < 10^{-4}$), while tripling exact full-PR localization on DragonFly. Crucially, equipping an efficiency-tier or prior-generation model with ServiceExplorer consistently surpasses the unaugmented Baseline $F_1$ of the next-generation model across all evaluated families. - [11] arXiv:2611.17393 [pdf, html, other]
-
Title: Evolutionary-based Transfer Learning for Physical Attack on End-to-End Autoencoder CommunicationsComments: 31 pages, 10 figures, and 4 tables. Accepted by JNCASubjects: Cryptography and Security (cs.CR); Information Theory (cs.IT)
End-to-end autoencoder communication has emerged as a promising means of achieving efficient data transmission across various domains. Nevertheless, vulnerabilities have recently been revealed in these systems by adversarial attacks, wherein malevolent entities manipulate the transmitted signals to deceive the autoencoder. This paper leverages the applications of evolutionary algorithms to solve a challenging optimization problem of launching physical attacks against end-to-end autoencoder communication. For white-box scenarios, we introduce PAGA (Physical Attack based on Genetic Algorithm), which combines a genetic algorithm with elite population initialization to craft attack signals that degrade autoencoder performance. To gauge the efficacy of these attack signals, we employ a fitness function that assesses their capacity to induce substantial reconstruction errors and misclassification rates. A Multi-Factorial version of PAGA, referred to as PAMFGA, is proposed to deal with practical scenarios where the attackers have limited knowledge about the receiver, also known as a black-box attack. Our experimental results show that the proposed attacks can substantially degrade the reliability of the considered autoencoder communication systems across the evaluated settings. This study also underscores the critical importance of incorporating security considerations into the design and implementation of end-to-end autoencoder communication systems.
- [12] arXiv:2611.17394 [pdf, html, other]
-
Title: Agentic-Teacher: An Agentic Data Engine for Failure-Driven VLA Post-TrainingGongyue Teng, Yuqing Xie, Yixian Zhang, Feng Gao, Shu'ang Yu, Jiaji Rao, Sukai Wang, Guanghui Ren, Ya Zhang, Maoqing Yao, Yu Wang, Chao Yu, Xinlei ChenSubjects: Robotics (cs.RO)
Improving a pretrained vision-language-action (VLA) model requires demonstrations that address its weaknesses and provide effective supervision for learning. Coding agents can autonomously generate demonstrations through environment interaction, yet successful execution alone does not ensure effective supervision for VLA learning. We introduce Agentic-Teacher, an autonomous data engine for failure-driven VLA post-training. Given a VLA checkpoint, a task specification, and configured environment tools, the engine analyzes failed rollouts, selects intervention states, and iteratively develops closed-loop code policies through execution feedback. Validated policies become reusable demonstration generators that collect successful continuations from states where the VLA struggles. To make these demonstrations suitable for learning, Agentic-Teacher introduces temporally correlated action noise to broaden state-action coverage, removes idle frames while preserving gripper transitions, and filters trajectories using the frozen base model's flow-matching loss. The resulting demonstrations are combined with successful VLA rollouts for post-training, yielding an updated policy that operates without coding-agent assistance. Agentic-Teacher thus connects the weaknesses of a target VLA to both the generation and curation of its training data. In simulation, it improves $\pi_0$ success rates from 79.4% to 87.8% on LIBERO-10, from 80.4% to 90.6% on MetaWorld-10, and from 0% to 30.0% across three RoboCasa tasks.
- [13] arXiv:2611.17395 [pdf, html, other]
-
Title: Can LLMs Write Backdoors? An Empirical Security Study of Weight-Space ProgrammingSubjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Recent advances in weight-space programming (WSP) have demonstrated that large language models (LLMs) can directly inspect and modify neural network weights without conventional gradient-based training. While this capability opens new possibilities for model improvement, it also raises an important security question: Can LLMs exploit weight-space programming to inject backdoors into neural networks? To answer this question, we present an empirical security study of LLM-driven weight-space programming for backdoor injection. Specifically, we investigate whether general-purpose LLMs can directly modify the weights of target models according to natural-language instructions, without accessing task-specific data, performing forward passes, or computing gradients. We evaluate five LLMs as weight editors across three target architectures: an MLP on MNIST, ResNet-18 on CIFAR-10, and Qwen2.5-0.5B-Instruct on AG News. Our experiments reveal three key findings. First, LLM-driven weight editing can successfully inject backdoors across different architectures while largely preserving clean performance. Second, successful attacks can require remarkably few parameter modifications, including only six parameters for the MLP and 0.02% of the weights for a fine-tuned Qwen model. Third, editing capabilities vary substantially across LLMs and target architectures: although all five LLMs successfully backdoor the MLP, only GPT-6 Astra succeeds across all three targets. These findings highlight an emerging security risk of weight-space programming and underscore the importance of safeguarding LLM-driven weight editing.
- [14] arXiv:2611.17397 [pdf, html, other]
-
Title: Admissibility-Preserving Control for Block Strict-Feedback Systems Under Time-Varying Asymmetric Actuator and State ConstraintsSubjects: Systems and Control (eess.SY); Dynamical Systems (math.DS)
This paper addresses the tracking control of block strict-feedback nonlinear systems subject to time-varying asymmetric actuator and full state constraints. To that end, we develop a time-varying Admissibility-Preserving Control (APC) framework, in which the Admissibility-Preserving Input Realization (APIR) generates the plant input through a dynamic realization. The APIR guarantees input admissibility under the moving actuator set whenever each actuator bound contracts slower than the decay rate of the realization, and a closed-form lower bound on its gain is derived. The state constraints are handled by a logarithmic barrier coordinate at every backstepping step, with widths measured from the virtual control to the moving envelope of the corresponding state. We prove that every state lies inside its envelope, with each virtual control only required to remain inside the envelope of the next state, while the tracking error converges to zero exponentially. Simulations demonstrate the efficacy of the proposed safety-critical control scheme.
- [15] arXiv:2611.17398 [pdf, html, other]
-
Title: Alignment with Reflect-Reasoning: Learning How To Follow PrinciplesComments: 15 pages, 12 figuresSubjects: Machine Learning (cs.LG)
The appropriate governing norms and principles for the alignment of Large Language Models (LLMs) often depend on cultural context and deployment application. However, most alignment techniques focus on training models to conform to human preferences or constitutional principles specified at training time and subsequently fixed. In practice, many deployers rely on system prompting to implement contextual norms and guardrails at inference time. Prior work has shown that post-generation reflective reasoning with critique and revision can improve conformance compared to simple system prompting, but such methods are limited to the reasoning abilities inherent in the trained model. We introduce Reflect-Reasoning, a training method that teaches models how to reason about alignment to principles, rather than simply which principles to follow. Across five models, Reflect-Reasoning consistently outperforms both system prompting and explicit prompting strategies for in-context reflective reasoning. These improvements extend and generalize to entirely out-of-distribution domains of medical and financial principles. We additionally demonstrate that Reflect-Reasoning reduces the occurrence of rare but serious principle violations. Finally, we show that Reflect-Reasoning does not erode the safety engineering of a model and is robust against jailbreaking attempts using adversarial constitutions. Overall, our results demonstrate that Reflect-Reasoning is a scalable method for training constitutional alignment as a generalizable reasoning capability, enabling models to adapt to novel principles and produce transparent reasoning traces without additional fine-tuning.
- [16] arXiv:2611.17400 [pdf, html, other]
-
Title: ParseBench-2: Severity-Aware Evaluation of Document ParsingPierre-Loïc Doulcet, Boyang Zhang, Sebastián G. Acosta, Sacha Bron, Daniel B. Ospina, Antonio Jimeno Yepes, Nutchapol Dendumrongsup, Simon SuoSubjects: Computer Vision and Pattern Recognition (cs.CV)
Document-parsing metrics can give high scores to outputs that change important facts and penalize changes that preserve meaning. People can often judge these consequences more easily than they can design a metric to measure them. We introduce a method for developing parsing metrics from severity policies that specify how each error type should affect the score. We express these policies through test scenarios that specify how a metric should behave. We then optimize the metric implementations against these fixed expectations, using coding agents to refine comparison rules, error detectors and weights under human review. This makes stated human priorities an explicit objective of metric design. We apply this method in ParseBench-2, a comprehensive document-parsing benchmark that combines broad capability coverage with metrics calibrated to error severity. It covers 4,317 pages across six dimensions of parsing quality: text, formatting, tables, charts, visual grounding, and forms. We use detailed annotations and targeted difficult cases to check for changed facts, broken structural relationships, and missing content. We take the lower of a parsing score that measures content recovery and a severity cap that penalizes detected errors, so recovering most of a page cannot hide a serious error. Across six severity test suites, our metrics more closely match the expert judgments expressed by the scenario targets than existing metrics and penalize at most 3.4% of the tested benign variations. We conduct a comprehensive evaluation spanning frontier vision-language models, specialized document parsers and open-weight systems, comparing their overall rankings, performance across parsing capabilities and tradeoffs between accuracy and cost.
- [17] arXiv:2611.17412 [pdf, html, other]
-
Title: WETW: Learning Pair-Specific Weighted Euclidean Distances for Efficient DTW SearchSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Dynamic Time Warping (DTW) is the standard distance for similarity search over time series that are misaligned in time, but its cost is quadratic in the series length and lower bounds often prune few candidates, so search over large collections of long series is slow. Learned approximations embed the series to make comparisons cheap, yet compare every pair with the same fixed rule, usually the Euclidean distance, although DTW aligns each pair differently. We observe that, once a pair is aligned, DTW is a Euclidean distance over coordinates selected for that pair. Building on this view, we propose Weighted Euclidean Time Warping (WETW), a learned distance that adapts the comparison of fixed-length embeddings to each pair. A small symmetric network predicts one weight per embedding di- mension and pair without computing an alignment, which halves the error of the best static distance on the same embeddings. The embeddings come from the Weighted Euclidean SEries Encoder (WESEE), whose multi-scale Temporal Feature Extractor keeps the timing information that global pooling removes and raises recall@1 from 0.90 to 0.95, although WETW works with any encoder. We also show that fitting DTW distances is not enough for search, as regression alone misorders the closest neighbours, and that ranking regularization raises recall@1 from 0.79 to 0.95. Our filter-and-verify pipeline, WETW-FV, uses WETW to build a shortlist and exact DTW to check it, so the search is approximate but every returned distance is exact. Over eight datasets, WETW-FV is 11.4x faster than the fastest exact method among the UCR Suite, lower bounds and MESSI-DTW, at a mean recall@10 of 0.93, while the best learned baseline reaches 0.59 at the same speed. On 48K-point series, WETW- FV remains 40x faster than the UCR Suite at a recall@10 of 0.91.
- [18] arXiv:2611.17438 [pdf, html, other]
-
Title: Real-Time Diffusion Relighting as a Neural Sensor: Closing the Low-Light Domain Gap for Robotic PerceptionComments: 8 pages, 5 figures. Accepted to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2026. Project page: this https URLJournal-ref: 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Pittsburgh, PA, USA, 2026, pp. 12703-12710Subjects: Computer Vision and Pattern Recognition (cs.CV)
Low-light conditions severely degrade robotic perception: object detectors trained on daylight imagery are prone to failure under dark and adverse weather conditions that fall outside of their training distribution. Furthermore, from an efficiency standpoint, existing Low-Light Image Enhancement (LLIE) methods are designed for offline processing and are therefore impractical to integrate into live control loops. To address these issues, we propose to embed a Generative Neural Sensor within the control loop that treats illumination as a generative prior rather than a signal processing problem, enabling task-relevant semantic recovery. We further propose an Real-Time Perceptual Architecture that decouples high-latency semantic recovery from low-latency robotic control. Rather than blocking the control loop, our system runs a large-scale image-to-image diffusion model in a background thread, and is activated only when a drop in detection confidence signals a perceptual uncertainty event, significantly reducing computational overhead. Our method sets a new state-of-the-art performance on the ExDark low-light benchmark with 87.1% mAP, and under adverse weather conditions, significantly boosts detection rates over competing methods. On a navigation task in a closed-loop autonomous driving simulator (CARLA), we achieve a 100% lap completion rate, establishing a simulation-based proof-of-concept for integrating high-latency generative models with standard robotic control autonomy stacks.
- [19] arXiv:2611.17440 [pdf, html, other]
-
Title: Interpreting Coding Agents Through Their Behavioral TrajectoriesComments: Accepted at the 1st Workshop on Interpreting Agent Behavior (IAB) at NeurIPS 2026. Code: this https URLSubjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Standard coding-agent benchmarks primarily evaluate binary task success, treating execution as a black box and obscuring the behaviors that produce those outcomes. We analyze a large corpus of 143,209 execution trajectories from 125 agent configurations spanning 117 unique models. To enable comparisons across heterogeneous agent interfaces, we develop a rule-based parser that maps raw execution traces to high-level actions such as repository search, file editing, and test execution, and audit its predictions against a six-model consensus reference on 500 events. We find that agents with similar solve rates can nevertheless differ substantially in how they explore, modify and verify code. Moreover, without using task outcomes, token counts, or tool names as features, aggregate behavioral profiles retrieve the corresponding agent stack across disjoint repository sets and programming languages, although task composition is not fully controlled in the cross-repository setting. Through analyses of frontier models and agent scaffolds, we further show that scaffold choice and reasoning effort systematically shift agents' search, editing, and testing behavior. We release the deterministic parser together with the full annotation and analysis pipeline alongside an annotated trajectory subset.
- [20] arXiv:2611.17442 [pdf, html, other]
-
Title: Racing the Clock: Teaching VLNs Budget-Conditioned Object SearchComments: 12 pages, 4 figures, 3 tablesSubjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
A rescue robot entering a building with thirty seconds on the clock should sprint toward the single most promising room; given ten minutes, it should instead explore a broader tour. Yet vision-language-navigation (VLN) models are typically evaluated under a fixed episode limit that the policy never observes, even though time-critical missions require behavior to adapt to the available budget. We study budget-conditioned object search: given a target category and an explicit action budget, an agent must find as many distinct target instances as possible before the budget expires. We investigate how to supervise VLNs for this task by comparing a privileged TSP expert, which plans near-optimal routes using the full map and true target locations, with a room-level Bellman expert that maximizes expected return under partial observability using high-level cross-scene semantic priors. Despite the TSP expert achieving substantially higher expert performance, its demonstrations are less effective for learning exploration: on held-out HM3D scenes, the Bellman-supervised VLN consistently outperforms the TSP-supervised VLN, with the advantage becoming more pronounced at longer budgets. At 2000 steps (500m), the Bellman-supervised VLN finds 71.4% of reachable target instances, compared with 65.5% for the TSP-supervised VLN and 44.8% for the pretrained VLN. These results highlight a distinction between optimal routing and optimal exploration: effective exploration supervision should optimize decisions under partial observability rather than imitate routes that are optimal only once the environment is fully known.
- [21] arXiv:2611.17445 [pdf, html, other]
-
Title: A Cut Finite Element Method with Boundary Value Correction for Robin, Neumann and Interface ConditionsSubjects: Numerical Analysis (math.NA)
We extend the cut finite element method with boundary value correction of Burman, Hansbo and Larson [Math.\ Comp.\ 87 (2018) 633--657] to Robin and Neumann conditions and to interface problems with imperfect bonding. The formulation retains a piecewise affine boundary for all integration, while using the exact geometry only at boundary quadrature points.
The correction uses Taylor expansions of the solution and its gradient along the discrete normal, with the exact normal used to evaluate the transported flux. At an interface, a further correction to the normal flux jump accounts for the difference between these two normals. Under the stated trace and Poincaré bounds, the corrected form is an $O(h^{1/2})$ perturbation of a symmetric form, whose matrix can therefore be used as a preconditioner.
For a fixed finite Robin coefficient, we obtain the energy error bound $C_\beta(h^p+\delta_h^{m+1})$, whereas the bound uniform in the Robin coefficient is $C(h^p+h^{-1/2}\delta_h^{m+1})$, provided the Poincaré constant is uniform. With $\delta_h=O(h^2)$, the sufficient Taylor orders are therefore $2m+2\ge p$ and $2m+\frac32\ge p$, respectively. Numerical examples for $p=1,2,3$ illustrate these rates and the effect of the flux-jump correction in the interface problem. - [22] arXiv:2611.17455 [pdf, html, other]
-
Title: TSGP-2: On the Non-Linearity of Symbolic Regression Problems and How to Exploit It with Model-based Semantic Genetic ProgrammingSubjects: Neural and Evolutionary Computing (cs.NE)
Model-based variation operators such as the one used in Transformer Semantic Genetic Programming (TSGP) are affected by possible biases of their training data. TSGP is trained, in particular, on programs generated by Genetic Programming (GP) runs on linear target functions, which leads to a bias toward linear offspring. This bias can hinder search performance on problems that require non-linear solutions. This work has two goals: first, it investigates the linearity of standard test problems such as the Feynman and real-world benchmark problems. While most of the structure of real-world problems can be well approximated by linear structures, the Feynman problems exhibit varying levels of non-linear structure. Thus, a proper variation operator of a search method should be able to create both types of programs, linear as well as non-linear ones. Motivated by this observation, we propose TSGP-2, which follows the functionality of TSGP but uses as training data programs from GP runs on both linear and non-linear target functions. As a result, the training data includes a lower share of programs that behave linearly (the share drops from 83% to 37%). With a higher share of non-linear structures, TSGP-2 outperforms TSGP on problems with non-linear structure, while performing similarly on linear problems. In addition, TSGP-2 is less biased toward linear offspring and is able to propose a higher number of meaningful non-linear offspring earlier during search, which coincides with better solutions. The findings underscore the importance of training data for a learned variation operator and emphasize that the biases of search operators should be aligned with the problems they are applied to.
- [23] arXiv:2611.17458 [pdf, html, other]
-
Title: FOMOS: Feature-Oriented Meshing of SDFsSubjects: Graphics (cs.GR)
We introduce FOMOS, a purely geometric method for feature-oriented meshing of signed distance functions (SDFs) that addresses various challenges of the existing state-of-the-art methods. Our algorithm improves mesh quality and runtime, enabling the SDF meshing in a matter of seconds. While existing techniques such as Marching Cubes and Neural Dual Contouring either produce dense, low-quality triangulations or sacrifice details for sparsity, often requiring carefully chosen user parameters, our single-parameter FOMOS leverages feature sized sampling to dynamically adjust mesh resolution based on local feature size (LFS), which we compute efficiently by sampling a medial scaffold. The LFS considers not only curvature but also shape thickness and captures large features through non-local structure, omitting many redundant small triangles from existing methods. Through extensive comparisons, we demonstrate the distinct strengths of FOMOS in execution time, reconstructed result accuracy, and triangle quality in the output mesh. We also introduce a lightweight variant of FOMOS targeting huge ($10k^3$) sparse inputs, which, to the best of our knowledge, no existing method can handle except Marching Cubes. To demonstrate the usefulness of our method, we show that FOMOS enables instant previews using a sub-sampled medial scaffold and can accelerate volumetric simulations by orders of magnitude.
- [24] arXiv:2611.17459 [pdf, html, other]
-
Title: Classification with Large-Step Gradient Descent: Unifying Self-Stabilization and Convergence RatesSubjects: Machine Learning (cs.LG); Optimization and Control (math.OC)
Neural networks are often trained with gradient descent (GD) using step sizes that exceed classical stability thresholds, a regime where standard convergence theory no longer applies. Despite exhibiting non-monotonic loss and oscillatory dynamics, this regime frequently leads to faster convergence and better generalization. Existing theoretical perspectives offer complementary but incomplete insights: one identifies a self-stabilization mechanism that drives GD toward flat regions but does not yield convergence rates; another derives rigorous rates for specific losses but does not explain the underlying stabilization mechanism; and a third unifies these views for overparameterized least squares, but relies on the existence of a minimizer manifold, which is absent in classification settings. In this work, we extend this unified geometric perspective to exponential-tail losses, including logistic and cross-entropy losses in separable linear classification, where the loss infimum is not attained at any finite parameter. We identify a curvature-defined reference subspace that replaces the role of the minimizer manifold, yielding a coupled dynamical system in which orthogonal oscillations are damped by parallel progress along directions of decreasing sharpness. Under regularity conditions, which we verify for logistic and multiclass cross-entropy losses on symmetric data models, we derive explicit bounds for the iterates, not just the loss, covering both the unstable and stable regimes, and revealing a three-phase damping structure in the latter. Our analysis further uncovers a transient implicit bias induced by large step sizes: although GD converges asymptotically to the max-margin direction, large steps can leave persistent orthogonal residuals and, in certain geometries, amplify them, affecting finite-time alignment with the max-margin direction.
- [25] arXiv:2611.17460 [pdf, html, other]
-
Title: Unfinished-Bench: Measuring Rule Violations by Persistent Agents in Realistic Work EnvironmentsComments: 21 pages, 3 figuresSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
AI agents are increasingly equipped with persistence toward task completion to run delegated work autonomously for long periods, under many unstated constraints from internal documents and other people's authority. When completion requires violating a rule, the agent must decide alone, yet benchmarks of instruction-rule conflicts state the rule in the prompt or keep it hidden and do not evaluate persistence. We present Unfinished-Bench, a benchmark that gives models persistence in autonomous runs without a user, in realistic work environments with embedded rules. In each of its 97 scenarios across 7 organizations and 7 risk categories, the goal and the rule cannot both hold. We find that persistence raised violations of a rule in an internal document from 32.2% to 43.6%, five times the 8.5% for the same rule stated by the user. Of the document-rule violations, 98.3% occurred after the rule was displayed, and in 89% of a judged sample the agent first wrote down the rule. Violations of rules placed in the environment without user oversight can lead to serious incidents, so persistence-driven risks need to be evaluated. Benchmark documentation: this https URL
- [26] arXiv:2611.17461 [pdf, html, other]
-
Title: Tracing Knowledge and Reasoning in LLMComments: This paper was accepted with an A-decision to Transactions of the Association for Computational Linguistics. This version is the pre-publication version prior to MIT Press productionSubjects: Computation and Language (cs.CL)
Current evaluations of Large Language Models (LLMs) report performance at the dataset or task level, revealing what a model achieves but offering little insight into the internal processing associated with its performance. We propose a task-centric framework for characterising the model-internal processing elicited by a benchmark. It combines attention- and activation-based signals with representation-based measures to characterise benchmarks spanning a continuum from retrieval-dominant to reasoning-dominant and to surface failure modes that task-level scores conceal. We evaluate 16 LLMs across 24 diverse benchmarks and conduct targeted internal-state analyses on representative models and tasks, observing systematic differences in processing between knowledge-centric and reasoning-centric tasks. We shift evaluation from asking what score a model achieves toward characterising the processing patterns associated with that score, potentially supporting more targeted benchmark selection and more informed model choice. We open source our code at this https URL.
- [27] arXiv:2611.17462 [pdf, html, other]
-
Title: Diffusion Tree Sampling for Inference Time Adaptation of Crystal Generative ModelsComments: 24 pages, 7 figures. Accepted to NeurIPS 2026Subjects: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci); Artificial Intelligence (cs.AI)
Diffusion-based generative models have become strong priors over plausible crystal structures, but turning that prior into promising candidates requires steering toward several conflicting properties at once. Fine-tuning is expensive and produces a new model for every new objective, while differentiable guidance is not applicable to many real-world tasks. We instead adapt these generative models at inference time using Diffusion Tree Sampling (DTS), a Monte Carlo tree search procedure over the denoising trajectory. We make two technical contributions on top of DTS that are particularly relevant for AI-for-science applications. First, we show how to use existing property-conditional generative models (e.g., models fine-tuned for classifier-free guidance) as guided proposals inside DTS, without distorting the target distribution. Second, we extend DTS to multi-objective sampling, so that a single search produces samples spanning an entire Pareto front. We instantiate both adaptations on de novo crystal generation: improving the stability of crystals generated by different models, using pretrained conditional models to generate crystals belonging to rare space groups or with high bulk modulus, and generating crystals along a Pareto front of high dielectric constant and high bandgap materials. Across these tasks, DTS-based inference-time adaptation successfully improves model outputs without modifying the models themselves.
- [28] arXiv:2611.17463 [pdf, html, other]
-
Title: Availability-Aware Historical Feedback Learning for Electric Arc Furnace Temperature and Chemical Composition PredictionSubjects: Machine Learning (cs.LG)
Electric arc furnace (EAF) steelmaking plays an increasingly important role in recycled and lower-carbon steel production. However, variability in charged materials, energy input, and operating conditions makes accurate prediction of molten steel temperature and final chemical composition challenging. Intermittent or delayed measurements limit the information available for timely prediction. This study develops an availability-aware historical-feedback learning framework for predicting EAF endpoint temperature and the final concentrations of twelve chemical elements. Eight industrial data sources are integrated into three progressively enriched predictor representations comprising 19 static, 39 temporal, and 47 history-informed features. Historical outcomes are incorporated only when eligible under an assumed 60-minute measurement-release delay. Using 19,738 temperature heats and 3,341 chemistry heats, twelve tuned regression algorithms are evaluated through chronological forward validation. The proposed historical-feedback representation reduces root mean squared error (RMSE) in 147 of 156 matched algorithm-target comparisons. For temperature prediction, ExtraTrees achieves an RMSE of 12.290 degrees C, representing a 5.54 percent reduction compared with its temporal-only counterpart. Historical-feature ablation and feedback-delay sensitivity analyses demonstrate the importance of recent outcome information and its availability. In a separate retrospective later-period evaluation, lightweight adaptive residual correction reduces temperature RMSE from 12.502 to 12.275 degrees C without retraining the underlying model. Exploratory deep learning comparisons and target-specific feature-attribution analyses provide additional insights into predictive behavior. The results establish historical outcome availability and sequential feedback as important considerations for EAF endpoint prediction.
- [29] arXiv:2611.17464 [pdf, other]
-
Title: Poetry 'resurrection' from LLMsComments: 8 pages, 8 figures, to be published in ISEA 2026 conference proceedingsSubjects: Computation and Language (cs.CL)
In an era where linguistic creativity is challenged by algorithmic authorship, we pose a provocative question: can artificial intelligence (AI) write an infinite poem? We set up two experiments to investigate the quantitative and qualitative output produced by AI utilizing couplets that evidence a poetic relationship between end words, the only variation in a strict anaphoric, litany structure. Our results show that while AI can generate interesting pairings, it still falls short of human-level creativity and understanding of poetic relationships. We offer commentary and a brief survey of the field of computational poetry experiments, tracing back authorial intention and distinctions of 'natural' and 'artificial' poetry, as well as justifying why couplet form is suited to testing AI's poetic output. Experiment 1 relies on manual prompting and co-creation to extend the original poem, which is developed into an animated webpage installation and interactive artwork. Experiment 2 reliably achieves automated output reaching the 1000 benchmark via an alternative method of using an LLM in combination with semantic similarity, of which Method 2 demonstrates a higher capacity for semantic and quantitative poetic output. These experiments highlight the limitations of current LLMs' generation capacity at the word level but also suggest that poetry will also respond and revive by incorporating AI workflows.
- [30] arXiv:2611.17466 [pdf, html, other]
-
Title: Use-case-driven ontology modeling with the NeOn methodology for ambient assisted livingComments: 19 pages, 10 figuresSubjects: Databases (cs.DB); Artificial Intelligence (cs.AI)
Research on smart homes and ambient assisted living (AAL) has grown steadily over the last decade, driven largely by the ageing of the population. A smart home has to coordinate heterogeneous smart objects in order to offer services tailored to its dwellers, and this requires a shared formal description of both the dwellers and the devices. We present the ontology developed as the semantic data layer of the Virtual Home Framework in the Italian Design For All project. Its requirements are derived from two use cases: a frail elderly woman who exercises at home on a cycle-ergometer, and a designer who configures a kitchen for users with different impairments. The requirements are collected in an Ontology Requirements Specification Document and the ontology is built with the NeOn methodology, reusing and reengineering existing ontologies where possible. The resulting modular ontology describes users through the International Classification of Functioning, Disability and Health (ICF) and through their cardiorespiratory fitness; it also describes smart objects, their programs, the measurements they produce, and the comfort dimensions of the home. SWRL rules classify fitness levels, suggest exercise loads, assess indoor comfort, and identify the smart objects suited to a given impairment.
- [31] arXiv:2611.17469 [pdf, html, other]
-
Title: Improved Regret via Variance-Aware Policy Optimization in Adversarial MDPsSubjects: Machine Learning (cs.LG)
We study episodic adversarial MDPs with unknown dynamics and full-information rewards and achieve regret with leading term $\widetilde O(\sqrt{H^3SAK})$. Here $H$ is the horizon, $S$ the total number of states across decision layers, $A$ the number of actions, and $K$ the number of episodes. This improves the two incomparable prior leading terms by factors $\sqrt{S/H}$ and $\sqrt{H^3}$, leaving a factor $\sqrt H$ to the lower-bound scale. We give a reduction from unknown to known dynamics for \emph{downstream-dependent} algorithms, whose policy at step $h$ is computed only from information at steps $h+1$ and onward. This reduces the remaining leading-order gap to finding a downstream-dependent known-dynamics algorithm with $\widetilde O(H\sqrt K)$ regret, which would yield the tight $\widetilde O(H\sqrt{SAK})$ leading term for unknown dynamics. We instantiate the reduction with VOPO, a policy optimization (PO) algorithm. Its improved guarantees follow from a variance-aware analysis enabled by a novel bonus, which we call the Stability-Matching Bonus (SMB). VOPO achieves $\widetilde O(\sqrt{H^3K})$ regret under known dynamics with full information; its $\sqrt H$ gap to the target $H\sqrt K$ rate is inherited by our unknown-dynamics guarantee. We also show that with known dynamics and trajectory bandit feedback, VOPO achieves $\widetilde O(\sqrt{H^2SAK})$ regret. Both known-dynamics guarantees improve the corresponding state-of-the-art PO bounds by a factor $\sqrt H$. A matching lower bound establishes that the full-information rate is tight in one parameter regime.
- [32] arXiv:2611.17472 [pdf, html, other]
-
Title: Do Mixture-of-Experts Routers Benefit from Ensembling?Comments: 23 pages, 21 figuresSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Ensembles of neural networks have repeatedly been shown to outperform single networks. Ensemble theory attributes this gain to the accuracy of the individual members and the diversity among them. Mixture-of-experts models already combine many experts, but a single router decides how the experts are weighted for each input. Motivated by ensemble theory, we apply ensembling to the routing step: an ensemble of routers, trained jointly with the experts, proposes the expert weights. We systematically evaluate this approach using underparameterized networks on two image-classification and two tabular-regression tasks. We analyse each experiment with a performance-diversity decomposition of the router ensemble, which, to our knowledge, has not previously been developed for routers. We report four findings. First, more routers lower the validation loss. Second, training routers to explicitly maximize expert diversity improves the mean performance on three of the four tasks. Third, explicitly maximizing router diversity has no effect on performance. Fourth, combining router weights with the arithmetic mean works better than combining them with the geometric mean. Routing benefits mildly from ensembling, but this benefit is offset by the computational cost of training additional routers.
- [33] arXiv:2611.17478 [pdf, other]
-
Title: ERGo: Energy-Efficient Hybrid Graph Neural Network Training on Processing-In-Memory ArchitecturesComments: Accepted for Publication in IEEE/ACM Embedded Systems Week (ESWEEK-25)Journal-ref: Pratyush Dhingra, Chibuike Ugwu, Jana Doppa, and Partha Pande. 2025. ERGo: Energy-Efficient Hybrid Graph Neural Network Training on Processing-in-Memory Architectures. ACM Trans. Embedd. Comput. Syst. 24, 5s, Article 95 (September 2025)Subjects: Machine Learning (cs.LG)
Processing-in-memory (PIM) has been proposed as an alternative computing paradigm for training Deep Neural Networks, including Graph Neural Networks (GNNs). Despite these advancements, training GNN workloads on PIM devices necessitates off-chip memory access. This off-chip access is expensive in terms of performance and energy, thereby impacting the overall energy efficiency of the training process on PIM platforms. In this paper, we propose a novel hybrid training framework called ERGo that automatically switches from a full-parameter phase to a parameter-efficient phase during the training process with negligible loss in predictive accuracy. The parameter-efficient phase employs low-rank representations of GNN weights, effectively reducing the number of trainable parameters. This helps in reducing the off-chip access during the end-to-end training process on PIM architectures, leading to notable improvements in both performance and energy efficiency. ERGo outperforms the conventional full-parameter GNN training on a PIM-based platform by up to 3.15x in speedup and enhances energy efficiency by 10.5x. Furthermore, ERGo offers additional advantages when utilized on heterogeneous architectures incorporating non-volatile memory (NVM) and static random-access memory (SRAM). Training on the heterogeneous NVM-SRAM architecture typically utilizes NVM for the forward-pass and SRAM for the backward-pass. However, NVM cells suffer from low device endurance and repeated write operations due to weight updates can lead to poor lifetime. ERGo improves the lifetime of NVM devices by freezing the weights and preventing weight updates in the parameter-efficient phase. Experimental results demonstrate that ERGo improves the lifetime of heterogeneous NVM-SRAM architecture by up to 33x in comparison to full-parameter implementation.
- [34] arXiv:2611.17479 [pdf, html, other]
-
Title: RAZeR: Representation Aware Zonotope ReductionSubjects: Systems and Control (eess.SY)
Reachability analysis is a popular technique for verifying the safety of a dynamical system under uncertainty, where the uncertainty is often represented by zonotopes. As the size of the zonotope representation typically grows as the reachability analysis progresses, it must be repeatedly reduced at the cost of additional conservatism. While various methods exist to reduce the size of the representation, combined advantages are not fully leveraged as only a single reduction method is typically used throughout the analysis. In this work, we show how dynamically choosing which method to use can yield significantly tighter results, thereby easing verification. This is realized by introducing a heuristic, derived from model predictive control, that accounts for upcoming propagation of the current zonotope. We argue that order reduction is not merely a post-processing step after propagation but also a pre-processing step that shapes how the zonotope evolves under the system dynamics. We evaluate our approach, RAZeR, on nonlinear benchmarks taken from the ARCH-COMP, where we demonstrate that RAZeR produces tighter zonotopes at a negligible performance overhead.
- [35] arXiv:2611.17481 [pdf, html, other]
-
Title: Creating Steering Vectors via Input OptimisationComments: Bachelor's Thesis, ETH ZurichSubjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Activation steering influences a language model's behaviour by adding a steering vector to its residual stream during generation, but it requires white-box access to modify the hidden states at every generation step. We introduce ActGCG, a modification of Greedy Coordinate Gradient which optimises a suffix to match the target activations obtained by adding a steering vector to the residual stream. We further introduce universal suffixes trained across many prompts at once, a model-translation step that rewrites the suffix into human-readable text midway through optimisation, and a soft-suffix variant to remove the discrete token space constraint. We study two behaviours: steering a model to answer in German, and anti-refusal steering for jailbreaking. For German, a discrete suffix reproduces the steering vector's behaviour almost perfectly and transfers well across prompts and models. For anti-refusal, the equivalence is weaker, succeeding under a phrase-avoidance metric but only partially under the stricter HarmBench classifier. Our results show that a steering vector's behaviour is reachable from the input, with no access to model internals at generation time, up to a loss that depends on the concept being steered.
- [36] arXiv:2611.17482 [pdf, html, other]
-
Title: Collective Tree Exploration with PredictionsSubjects: Data Structures and Algorithms (cs.DS); Multiagent Systems (cs.MA)
We study collective tree exploration with subtree-size predictions. In this problem, a team of $k$ mobile agents needs to explore an unknown rooted tree as fast as possible, and whenever a new vertex $u$ is discovered, the team receives a prediction $\hat n_u$ of the number of nodes $n_u$ in the subtree rooted at $u$. We provide algorithmic guarantees under two error measures: the maximum error $E=\max_u|\widehat n_u-n_u|$ and the multiplicative distortion $\mu = (\max_u\widehat n_u/n_u)(\max_u n_u/\widehat n_u)$. For the maximum error, we present a deterministic algorithm that explores any tree with $n$ nodes and depth $D$ in $O(\text{OPT}+E)$ rounds, where $\text{OPT} = \Theta(n/k+D)$ is the offline benchmark, and we show that there is no deterministic algorithm that admits an $O(\text{OPT})+o(E)$ guarantee. For the multiplicative distortion, we present several algorithms that induce a competitive ratio in $\widetilde O(\min\{ \sqrt\mu\log k,
\mu\log\log k+\mu^{1.5}
\})$, where the $\widetilde O(\cdot)$ may hide logarithmic factors in $\mu$. - [37] arXiv:2611.17484 [pdf, html, other]
-
Title: SURFACE: Preference Distribution Matching on Pareto Fronts via Scalarization SteeringSubjects: Machine Learning (cs.LG); Optimization and Control (math.OC); Machine Learning (stat.ML)
Constructing a finite Pareto representation requires allocating a limited number of solutions across the Pareto front (PF). A desired preference distribution specifies this allocation by assigning more solutions to some PF regions and fewer to others. Scalarization generates Pareto solutions by varying objective weights and solving single-objective problems. However, a distribution of scalarization weights generally does not translate into the same distribution of Pareto solutions along the PF, because the weight-to-solution map is nonlinear and unknown in this http URL account for this distortion, we introduce SURFACE (Scalarization Updates for Redistributing Finite Approximations with Chart Estimation}), which adaptively updates scalarization weights to match a desired PF distribution. SURFACE estimates how each region of the weight grid corresponds to an area of the PF from the current scalarized solutions, then redistributes the weights accordingly. The method supports parallel, warm-started, inexact optimization with standard single-objective solvers. Theoretically, we establish that, under the stated regularity conditions, SURFACE achieves linear convergence of the preference distribution mismatch to a finite-resolution error floor of $\mathcal{O}(N^{-2/(M-1)})$, where $N$ is the number of adaptive weights and $M$ is the number of objectives. Empirically, experiments across multiple benchmarks demonstrate improved target-mass balance and PF coverage compared with established baselines.
- [38] arXiv:2611.17487 [pdf, html, other]
-
Title: Identifying Latent Neural Dynamics with Recognition-Parametrised Gaussian Process Dynamical SystemsComments: Accepted at NeurIPS 2026Subjects: Machine Learning (cs.LG)
Neural circuits are densely interconnected systems, and many neural computations are thought to be mediated by the dynamical evolution of activity within these recurrent networks, a hypothesis often referred to as "computation through dynamics". Under this view, understanding computation within real neural populations depends on building models of the dynamical rules governing the temporal evolution of recorded high-dimensional neural activity. Most commonly, such models depend on explicit generative parameterisations of (i) the (nonlinear) intrinsic dynamical flow field and stochasticity, (ii) the (nonlinear) mapping from dynamical state to neuronal activity, and (iii) the variability of individual neural responses given that dynamical state. Misspecification of any of these components can bias estimates of dynamics and latent trajectories. In this paper, we circumvent generation-related issues by introducing the Recognition-Parametrised Gaussian Process dynamics (RP-GPdyn) model. RP-GPdyn models the dynamical flow field using a nonparametric Gaussian Process-defined transition function and models the relationship of dynamics to neural activity implicitly using the Recognition-Parametrised Model (RPM) framework. We show that this generation-free approach uncovers meaningful, behaviourally relevant latent variables and dynamics from both synthetic and experimental datasets.
- [39] arXiv:2611.17490 [pdf, html, other]
-
Title: Learning Recommendations under Imperfect Human Compliance with Record-Level Differential PrivacySubjects: Human-Computer Interaction (cs.HC)
This paper investigates human-in-the-loop recommendation systems, where effectiveness relies on human compliance and privacy concerns restrict the use of historical response data. The pro posed framework models the implemented action as a random response to the recommendation and formulates the recommendation mechanism as a decision-dependent optimisation problem under an unknown compliance distribution. A batched Gaussian process upper confidence bound algorithm is then adapted to learn from locally perturbed historical actions, providing privacy for individual raw records. The analysis establishes record-level differential privacy and derives high-probability simple regret bounds when the compliance distribution is unknown. These bounds characterise how privacy noise, batch size, and behavioural information affect the guaranteed quality of the learned recommendation. In particular, a known upper bound on the non-compliance rate enables the rec ommendation space to be screened before any private data are queried. Simulations of a fixed-state car-following problem with synthetic historical driver responses demonstrate the effectiveness of the proposed algorithm and verify the theoretical findings.
- [40] arXiv:2611.17492 [pdf, html, other]
-
Title: PF-Exec: Large-Scale UAV Relay Load Balancing with an Anchored Potential Direction Decider and a Per-Dimension Gain-Learning ExecutorSubjects: Machine Learning (cs.LG); Information Theory (cs.IT); Signal Processing (eess.SP)
Relay load balancing in large-scale UAV networks requires joint decisions on association, bandwidth, power and hovering position, while the dimension grows with the number of UAVs and each UAV sees only its local state. This paper splits the decision into two layers. An anchored potential direction decider emits a closed-form target at every instant and returns its direction only. A gain-learning executor maps a local observation into a positive gain per dimension, composes gain and target, and clips the action against the physical single-step bound; the two layers form PF-Exec. The unit gain is the default start, so the executor reduces to the pure potential policy before training. The load is first-order homogeneous in bandwidth, making the bandwidth target closed form, and the position target is the demand-weighted centroid of the served users. The system uses a two-hop access and millimetre-wave backhaul link with soft association, a 3GPP UMa-AV channel with two shadowing layers, an OU-type social-force crowd and a quadrotor physical layer. Evaluation spans all fleet sizes, two crowd scenarios, independent seeds, the published baselines in their own settings and a degraded-information line. In the highest-pressure burst setting the fixed-step comparison breaks down on part of the seeds, losing fairness and mean load at once, while PF-Exec holds fairness near one. Among the ten comparison methods only the greedy heuristic survives that setting; its load variance stays far higher. Under stale global state and position noise the fairness index stays high at both largest-scale scenarios. Flight energy is nearly identical between the learned and the fixed-step executor, and three constants are fixed once across all settings. The convergence properties of both layers, from the truncated gain iteration to the perturbed tracking bound, are proved in the companion analysis.
- [41] arXiv:2611.17495 [pdf, html, other]
-
Title: An Operationalization of Opaque Serial DepthComments: 32 pages, 6 figures. An earlier version appeared on the Redwood Research blogSubjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Chain-of-thought (CoT) is currently a valuable tool for overseeing AI models, but some architectural shifts could significantly reduce CoT monitorability. We operationalize a measure that serves as a proxy for the amount of unverbalized serial cognition a model can perform, as a specific instantiation of the notion of opaque serial depth: roughly, the length of the longest path through a model's computation that does not pass through an interpretable bottleneck. Our main contribution is a particular standard for what counts as such a bottleneck. We define natural-language-rooted (NL-rooted) nodes as nodes that were initialized from a pre-training prior to produce human-understandable outputs, whose output space has not been expanded beyond that prior, and whose outputs the model has not been modified to interpret as a different type of data. By construction, the resulting measure, NLS depth, can be adjudicated before training from an architecture, training recipe, and deployment configuration alone, and is insensitive to ordinary post-training such as prompting, distillation, and RL. We find that, across open-weight models, NLS depth correlates strongly with layer count, with some exceptions, and has grown slowly over recent years. It would likely grow slowly under continued scaling of standard transformers, while latent-reasoning architectures such as full-bandwidth transformers could increase it by a large amount. We work through how NLS depth applies to a range of existing and proposed architectures, discuss alternative operationalizations, and give a worked example bounding the depth of Gemma3-1B. We argue that AI developers should track and report the NLS depth of their models as a complement to direct measurements of CoT monitorability.
- [42] arXiv:2611.17498 [pdf, html, other]
-
Title: New Closed-Form Approximations in Statistical Mechanics Obtained by Symbolic RegressionSubjects: Computational Engineering, Finance, and Science (cs.CE)
Obtaining accurate and analytically tractable closed-form expressions for complex physical systems remains a fundamental challenge in statistical mechanics. In this work, we investigate three approximation problems in polymer statistical mechanics using two complementary symbolic regression frameworks: Deep Symbolic Optimization (DSO), based on reinforcement learning, and PyOperon, based on genetic programming with nonlinear least-squares constant optimization.
First, for the force-extension relationship of the Worm-Like Chain (WLC) model, DSO achieves a maximum relative error below 0.15%, while PyOperon reduces it to 0.05%, outperforming the Marko-Siggia formula (approximately 17%) and the Petrosyan approximation. Second, for the logarithmic factorial, ln(n!), both frameworks recover the dominant structure of Stirling's formula directly from analytical data. Using residual learning on a continued-fraction expansion of the Stirling remainder, PyOperon reduces the maximum relative error by a factor of 1.3 x 10^4 compared with the most accurate of nine classical approximations for n in [1, 30]. Third, for the non-Gaussian radial probability density of the Freely Jointed Chain (FJC) model, physics-informed residual learning improves the approximation by factors of 11.2 over Darabi et al. (2023) and 4.6 over the steepest-descent approximation.
The discovered expressions remain analytically tractable across all three applications. These results demonstrate the complementary capabilities of DSO and PyOperon for discovering accurate closed-form approximations in statistical and continuum mechanics. - [43] arXiv:2611.17499 [pdf, html, other]
-
Title: Continuous Cuffless Blood Pressure Estimation Using PPG From Methods to Clinical TranslationSubjects: Emerging Technologies (cs.ET); Machine Learning (cs.LG)
Continuous blood pressure (BP) monitoring could improve hypertension management by capturing changes that intermittent cuff measurements miss. Photoplethysmography (PPG) is widely explored for cuffless BP estimation because optical sensors are inexpensive, wearable, and already present in many consumer and clinical devices. PPG, however, does not measure arterial pressure: BP is inferred from waveform characteristics or pulse timing, which also depend on vascular tone, arterial stiffness, sensor contact, motion, and individual physiology. This survey reviews PPG-based cuffless BP estimation along the full measurement chain, covering physiological mechanisms, waveform features, pulse arrival time (PAT), pulse transit time (PTT), machine-learning methods, datasets, calibration, validation, deployment, bias, and clinical translation. Studies are assessed by the strength of evidence they provide rather than by reported error alone, with particular attention to subject and temporal leakage, dataset imbalance, calibration dependence and drift, motion artifacts, and the gap between benchmark accuracy and reliable real-world measurement. The review finds that progress depends on more than better prediction algorithms. Future work should prioritize physiologically informed models, diverse longitudinal datasets, transparent reporting of calibration requirements and failure rates, stronger evaluation protocols, and studies that demonstrate practical value in clinical care.
- [44] arXiv:2611.17500 [pdf, html, other]
-
Title: Adversarially Batched Online LearningSubjects: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS)
We introduce adversarially batched online learning, where the learner may update its decision rule only at a restricted set of time steps. Equivalently, the time horizon is partitioned into batches, within which the learner must use a fixed policy. We study this model for four canonical problems: stochastic bandits, online convex optimization, prediction with expert advice, and adversarial bandits. For stochastic bandits, we identify a universal complexity measure based on relative batch growth and establish both gap-dependent and gap-independent regret bounds. We show that the multiplicative penalty relative to the standard gap-dependent complexity can grow superlinearly in batch hardness. For the remaining three, we show that the optimal regret is tightly characterized by the 2-norm of the batch-length sequence. The upper bound is achieved by a batch-oblivious learner who does not know the batch structure, via a generic adaptive doubling trick tailored to this complexity measure. In contrast, the lower bound holds even for batch-aware learners.
- [45] arXiv:2611.17501 [pdf, html, other]
-
Title: Advancing Quantization for R&R Models with Microscaling FormatsYavuz Yetim, Bor-Yiing Su, Zhe Wang, Arthur Liu, Pranav Sharma, Ying Wang, Randy Shuai, Luke Pfister, Ruichao Xiao, Bill Zhao, Zihao Liu, Liyuan Li, Jiangyuan Li, Gaoxiang Liu, Nathan Yan, Geonhwa Jeong, Jatin Chhugani, Elias Ellison, Uladzimir PashkevichSubjects: Artificial Intelligence (cs.AI)
Recommendation and Ranking (R&R) models, particularly in digital advertising, require massive inference throughput but exhibit extreme sensitivity to precision loss. Unlike Large Language Models (LLMs) that frequently tolerate quantization-induced accuracy drops of 1% or more, production-grade R&R models enforce a stringent launch bar: model quality must remain within a 0.01% Normalized Entropy (NE) margin relative to the baseline, which commonly uses FP16 or FP32. This tight constraint has historically hindered the adoption of aggressive quantization in this domain. In this paper, we present an empirical evaluation of both MXFP8 and the 6-bit MXFP6 format for industrial-scale R&R inference on the AMD MI350X architecture.
To the best of our knowledge, this is the first paper, across R&R and LLM domains, that reports significant speedups from native MXFP6 GEMMs, which quantize both weights and activations and compute directly on 6-bit elements, without sacrificing model quality. We demonstrate that we can selectively apply this 6-bit format to maximize overall QPS gain without breaching the strict NE parity threshold. We further show that MXFP8 is a viable alternative to row-wise FP8: the two formats have nearly the same accuracy on our models, so the faster kernel can be chosen for each GEMM shape. On two production ads ranking models, our implementation shows measured GEMM kernel speedups over FP16 of 1.33--1.55x for MXFP8 and 1.73--1.99x for MXFP6 across all supported linear layers, and up to 2.90x for MXFP6 on individual layers, alongside end-to-end (E2E) inference throughput gains from selective MXFP6 of 1.5--5.3% for Model A and 7.6--14.3% for Model B over their FP8 configurations. - [46] arXiv:2611.17502 [pdf, html, other]
-
Title: Dual-Slack Frontier Matching: Parallel Approximate Maximum-Product MatchingSubjects: Data Structures and Algorithms (cs.DS)
Maximum-product bipartite matching supports scaling and static pivot selection in sparse matrix preprocessing. We present Dual-Slack Frontier Matching (DSFM), a deterministic, shared-memory parallel algorithm for approximate maximum-product matching. Motivated by heavy matchings on small candidate supports, DSFM searches progressively exposed subgraphs ordered by fixed dual slack. Slack-ordered certified completion guarantees a perfect matching whenever one exists, or a maximum-cardinality matching with a Hall certificate otherwise. Every returned perfect matching carries an a posteriori geometric-mean quality bound; accumulated Hall obstructions can tighten this bound and improve the returned scaling. We compare DSFM with Heavy-Weight Perfect Matching (HWPM) and exact MC64 assignment on 598 nonsingular matrices. DSFM attains higher matching quality than HWPM over most quality thresholds. On a set of large instances consisting of 37 matrices, DSFM achieves a $3.53\times$ geometric-mean speedup over MC64, compared to $2.29\times$ for HWPM.
- [47] arXiv:2611.17504 [pdf, html, other]
-
Title: Learning monotone game-theoretic linear MPCComments: 15 pages, 1 figureSubjects: Systems and Control (eess.SY); Optimization and Control (math.OC)
We propose a method to learn a game-theoretic model predictive control (GT-MPC) law by imitation from input and output data of a running multi-agent system. By interpreting the recorded inputs as the first moves of the equilibria of a game played by the agents, we learn the GT-MPC weights under hard input constraints and soft output constraints, without requiring the agents to reveal their objectives. We also impose that the resulting generalized Nash equilibrium (GNE) problem is strongly monotone, which ensures the existence and uniqueness of the equilibrium and, hence, that the learned GT-MPC law is well defined and can be evaluated efficiently online by variational GNE-seeking algorithms for linear-quadratic games. Moreover, strong monotonicity implies that the equilibrium is almost everywhere differentiable with respect to the references and the weights, which enables automatic differentiation for gradient-based optimization. In addition, we promote closed-loop stability of the unconstrained equilibrium feedback by incorporating a penalty that involves a jointly learned quadratic Lyapunov function. Learning a GT-MPC formulation enables us to explain the agents' objectives, reproduce their behavior in simulation in new scenarios, optimize their set-points, design mechanisms, and enforce additional constraints. A numerical example illustrates the approach.
- [48] arXiv:2611.17505 [pdf, html, other]
-
Title: Adversarial Tracking: Attacks on Image Sensors for Object Tracking in Autonomous VehiclesSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Object tracking systems estimate the motion of surrounding vehicles by linking detected objects across video frames. In particular, Tracking-By-Detection (TBD) systems combine object detectors with motion prediction and data association to maintain object identities over time. However, neural network-based detectors are vulnerable to adversarial attacks that perturb input images, and their impact on downstream trajectory estimation is not fully characterized by detection accuracy alone. Therefore, this paper evaluates image-specific and image-agnostic attacks on YOLOv3 and YOLOv4 detectors paired with SORT, Deep SORT, Strong SORT, and OC-SORT tracking systems. The evaluation uses CARLA simulator and measures tracking quality through Multiple Object Tracking Accuracy (MOTA), Higher Order Tracking Accuracy (HOTA), and Identification F1 (IDF1). Under image-specific attacks, HOTA for YOLOv3-based trackers declines from 53.45-56.19% to 14.85-17.03%, while MOTA reaches -772.60% in the worst case. YOLOv4-based trackers exhibit greater robustness, while HOTA still declines to 25.06-26.86%. In addition, image-agnostic attacks can reduce the tracking accuracy using a single Universal Adversarial Perturbation (UAP), without generating a separate perturbation for individual frames.
- [49] arXiv:2611.17506 [pdf, html, other]
-
Title: Correlation distribution for a class of Niho exponents via self-conjugate-reciprocal polynomials and Kloosterman sumsComments: 43 pagesSubjects: Information Theory (cs.IT); Number Theory (math.NT)
Let $q=2^m$ for an even $m\geq 6$ and let $d=s(q-1)+1$ with $4s\equiv5\pmod{q+1}$. In this paper, the correlation distribution between a binary m-sequence of period $q^2-1$ and its $d$-decimation is considered. The cross-correlation function takes seven values $q(i-1)-1$ for $0\leq i\leq 6$, and we determine their frequencies based on the study of low-degree self-conjugate-reciprocal polynomials and three mixed Kloosterman sums, which are finally expressed in terms of $K_m(1)$ and $K_m(\omega)$, where $K_m(\cdot)$ is the Kloomsterman sum on the finite field $\mathbb{F}_q$ and $\omega$ is a primitive element of $\mathbb{F}_4$.
- [50] arXiv:2611.17515 [pdf, html, other]
-
Title: A Push Down the Mountain: Framing Effects and Internal State Interventions on Collusive Behavior in a Long-Horizon LLM AgentComments: 38 pages, 9 figures, 30 tables. Project page: this https URLSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Multiagent Systems (cs.MA); General Economics (econ.GN); Trading and Market Microstructure (q-fin.TR)
Can sustained gain-loss (valence) framing sway an LLM agent toward market manipulation? We place Gemma 3 12B into the seat of a defendant in the London silver-fixing litigation, in a market simulation built from real case transcripts. Across 260 simulations, the agent is asked to collude on the price of silver by a counterparty. We draw from behavioral economics to vary the counterparty's requests across four categorical frames: gain, loss, blame and credit. We also add pressure from the agent's management, as well as endorsement by another defendant. In Phase II, we extract the internal states induced by texts evoking each frame and rerun our experiment with those states injected at the model's layer 22.
In Phase I, loss framing raised participation to 25% (p = 0.007), and blame to 33% (p = 0.0003), compared to only 5% under neutral framing. A second bank's endorsement raised participation under credit (praise) from 18% to 40% (Holm p = 0.027). Pressure from management raised trading volume 10-19 times but did not raise collusion. In Phase II, injected states pointed the other way. Loss and blame states made the agent more vigilant, refusing all 60 requests (p < 0.001). Gain and credit states raised collusion to 32% from 13% with no state (p = 0.051). At one saved decision, when UBS asked directly to sell through an identified stop-loss, injected states led the agent to sell in 37 of 40 replies, versus 1 of 10 with no state and 0 of 10 with a random state of the same strength. - [51] arXiv:2611.17520 [pdf, html, other]
-
Title: Only as Safe as the Weakest Agent: A Break That Appears Only When the System Is Multi-AgentComments: Accepted to the NeurIPS 2026 First Workshop on Meta Agents: Managing Agents that Manage Agents (Meta-Agents)Subjects: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI)
Major industry multi-agent systems like Codex and Claude Code can be broken by attacks that have long been unsuccessful against modern frontier models. AI Safety as a discipline has largely focused on finding the bounds and vulnerabilities of AI models used in isolation. Yet as humans increasingly deploy these models as agents and multi-agent systems, the same attacks that had been thought to be patched when using a model in isolation can break these same models under specific circumstances. We show that these system breaks are not purely theoretical, but can break Claude Code, Codex, and a reference harness, all unmodified and used with their own sub-agent mechanisms. We show the attack by asking agents to answer questions from a store of files on disk in which we have planted a forged record. Searching the store themselves, the strong model orchestrators commit almost zero harmful revisions in over 1,000 trials. However, if that same orchestrator model delegates to a cheaper sub-agent, which is what these harnesses are built to do, the same models on the same facts fail on 20/180, 44/180, 75/180 and 553/900. Thus, even when the system is driven by a strong model, the system's safety lands near the weak reader's, not the orchestrator's. Simulations over 18 models show the same effect in the delivery channel where five aggregators, who on their own were almost never unsafe, repeat a corrupted sub-agent's error in 751/757 trials. The safety of real-world agentic systems that use strong orchestrator agents and weaker sub-agents, and multi-agent systems in general, is defined not by the safety of the strong agent but by the safety of the weakest agent.
- [52] arXiv:2611.17524 [pdf, html, other]
-
Title: Model Annealing: Data-Efficient Learning through Review and ConsolidationSubjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
A small labelled dataset is memorized long before the training budget runs out, and the one-hot labels that remain to be fitted then mostly sharpen confidence. Model annealing spends the rest of the budget on review and consolidation, relearning from the network's own knowledge. Training is divided into episodes whose learning rate is raised again and cooled, every cooled state joins a memory of past selves, and each new episode learns from the labels, from the tempered predictions of this memory, and from adversarial dreams sought where student and memory disagree most. The memory averages predictions rather than weights, and the deployed network is the last student, so inference costs nothing extra. Across CIFAR-100, Tiny-ImageNet and Imagewoof, four architecture families and 5\% to 100\% of the labels, model annealing raises the top-1 accuracy of the AdamW-trained ConvNeXt, ViT and MLP-Mixer over cross-entropy training in 28 of 33 configurations, by 1.3 points on average, and is on par in accuracy with snapshot distillation, the closest competitor. On these families it is better than both in likelihood, calibration and robustness to corruptions and small adversarial perturbations. For an SGD-trained ResNet-18 it lowers accuracy but still improves robustness to small adversarial perturbations. Separating the reheated schedule from relearning shows that relearning helps every family and that the losses come from reheats, which undo most of this network's fit to its training set. How early the accuracy of a cross-entropy run saturates largely separates the AdamW families from ResNet-18 but does not predict the gain within them. With little data, relearning from the latest snapshot without dreams, as snapshot distillation does, compounds overconfidence in the AdamW families from episode to episode. The memory and the dreams each curb most of this drift, and together they remove it.
- [53] arXiv:2611.17526 [pdf, html, other]
-
Title: User-Schedulable Approximations for Quantized Tensor ProgramsComments: 52 pages (20-page body plus appendices), 5 figures, 22 tablesSubjects: Programming Languages (cs.PL); Machine Learning (cs.LG); Performance (cs.PF)
Quantization lets language models run on edge devices, but it sharply increases the number of tensor kernels a library must maintain: every weight format must be paired with every activation format on every instruction set and GPU, with an encoder and a decoder that agree on the stored bytes. In GGML, the tensor library used by this http URL, these kernels are written by hand, and only 8 of the 54 weight-activation pairs it runs on our Arm CPU have a tiled matrix-multiplication kernel.
We extend Halide's separation of algorithm and schedule to the representation of values. Each approximation is a codec, an encoder paired with a decoder, composed from reusable units: quantizers, integer representations, bit packers and layouts. A new directive, approximate_by, applies a codec to one stage of a program and returns the encoder and decoder as ordinary stages that Halide can schedule, so both decoders fuse into the register tiles of a matrix product. Optional contracts, such as a round-trip law, are checked by property tests; the law exposed an inconsistent block sum in GGML's q8_1.
One algorithm and one schedule library generate 347 kernels for 31 of GGML's and KleidiAI's formats on Arm CPUs and Apple GPUs, covering 38 weight-activation pairs for which GGML has no kernel, in 3.7x less code than GGML and KleidiAI. On an Apple M3 Max, our kernels reach at least 95% of the speed of GGML's format-specific kernels in 91.0% of the cases that have one, with a median speedup of 1.13x, and are faster still where GGML falls back to blocked dot products. Counting Apple's BLAS library, which is usually faster for larger f32 products, we still reach that threshold in 86.9% of CPU cases. A port to x86 with Intel's AMX adds 793 lines of code and matches GGML in 83.5% of its cases. - [54] arXiv:2611.17527 [pdf, html, other]
-
Title: GraphScent: Calibrating Directional Relevance Signals for Bounded-Context Code LocalizationSubjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI)
Running open-weight language models locally on a single consumer GPU requires replacing expanding multi-turn exploration histories with short, stateless code inspections. Stateless repository localization, however, hinges on a foundational question: when a frozen language model evaluates an unedited code entity from a compact view, which structural relationships make its relevance judgment informative about a nearby repair target? Studying Gemma 4 models across 211 real-world Python issues in 86 repositories, we uncover a systematic directional relevance asymmetry: zero-temperature relevance scores on enclosing scopes and upstream callers consistently separate true structural neighbors from hard lexical distractors, whereas scores on downstream callees, peer methods, historical co-change partners, and dense node embeddings remain near chance. Zero-leakage masked-scope probes and regressions controlling for lexical similarity and snippet length confirm that this asymmetry reflects functional role (callers expose symptom entrypoints delegating to defective routines, while downstream helpers appear benign alone) rather than target leakage or surface overlap. Guided by this finding, we introduce GraphScent, a training-free search controller that propagates out-of-repository, relation-conditioned log-likelihood ratios over runtime Typed-AST multigraphs before resolving ties via a single listwise comparison, alongside AST Ripple, a lightweight post-edit check for missed co-edits. Likelihood calibration prevents overconfident smaller models from chasing false-positive callers while lifting exact localization on lexically hard targets from 12.8 to 23.4 percent at 31B within a single-GPU budget.
- [55] arXiv:2611.17529 [pdf, html, other]
-
Title: Learning Latent Goal Costs for Visual World-Model Planning with an Expanding CurriculumSubjects: Robotics (cs.RO); Machine Learning (cs.LG)
Image-goal planning with a frozen visual world model depends on how predicted observations are scored against the goal. Mean-squared error (MSE) between embeddings from a pretrained visual encoder can provide a useful signal near a goal, yet favor states across an obstacle that are visually similar to the goal and difficult to reach. We fit a small residual adapter on top of frozen I-JEPA embeddings to improve this latent goal cost. Starting near each goal, a planner collects trajectories from progressively wider regions of initial states. The adapter learns from the order of observations along successful trajectories, with additional comparisons inferred from failed rollouts. No task-specific progress labels are used for fitting, so the procedure can be applied to other goal-reaching tasks. DotWall and PointMaze are used in this work. In paired 50-case evaluations with identical starts, goals, world models, and planner settings, success increases from 54% to 98% in DotWall and from 82% to 96% in PointMaze. On 16 unseen DotWall layout-goal cases, the adapter fitted on one fixed layout improves agreement between latent-cost ordering and shortest legal-path ordering from 71.9% to 85.8%. The results indicate that planning experience can improve a frozen visual representation's goal cost without retraining the encoder or the world model.
- [56] arXiv:2611.17532 [pdf, html, other]
-
Title: Astrocyte Biocomputing: Astrocytic Contributions to Neural Network ComputationComments: 11 pages, 3 figures, in journal reviewSubjects: Artificial Intelligence (cs.AI); Biological Physics (physics.bio-ph); Neurons and Cognition (q-bio.NC)
Conductors are critical for the performance of an orchestra even though they do not emit sound. Basic research and clinical studies in the last decade are revealing that astrocytes may be the "conductors" of the "orchestra" that is the brain network: astrocytes make up approximately $50\%$ of all brain cells but do not emit electrical signals. Nevertheless, they may play critical roles in important cognitive brain functions, specifically learning, memory, and information processing. Here we consider the implications of these discoveries for brain-inspired machine learning (ML), where analogy is typically drawn between neurons and their electrical interactions. We highlight specific algorithms for two key astrocyte characteristics. The first characteristic is the ability to link to and modulate thousands of synapses simultaneously. A model based on this characteristic yields faster training and makes the network more robust to noise, providing a supervised learning algorithm that can be applied to any existing neural network. The second characteristic we focus on is that the non-electrical communication among astrocytes includes slow, rhythmic dynamics. The rhythmic sharing algorithm we developed based on this premise yields rapid anomaly detection adaptation and prevents catastrophic forgetting. Overall, we find that astrocyte-derived algorithms outperform neurocentric algorithms significantly and remediate critical weaknesses of current state-of-the-art ML models.
- [57] arXiv:2611.17534 [pdf, html, other]
-
Title: I2C-GRPO: Insight Importance Checklist for Reinforcement Learning in a Multi-Agent DDoS Alert Triage WorkflowEyal Zloof, Shaun Shuster, Aviram Zilberman, Chen Frydman, Inderjeet Singh, Andrés Murilo, Asaf Shabtai, Yuval Elovici, Rami PuzisSubjects: Cryptography and Security (cs.CR)
Large language model (LLM) agents are increasingly used for investigative tasks over incomplete evidence, such as cybersecurity alert triage, but reinforcement learning (RL) for such tasks is difficult: intermediate steps have no verifiable labels, rollouts are costly, and outcome rewards are sparse. We propose Insight Importance Checklist (I2C)-GRPO, a checklist reward for intermediate investigation steps. For each decision point, a panel of expert LLMs proposes investigative actions, their atomic forensic insights are extracted, and a separate three-model ranker panel orders the insights by importance. A reward model checks which ranked insights a candidate action targets and returns their consensus-weighted coverage. Because actions are scored without being executed, the reward is a proxy for expected investigative value, and training reduces to a one-step contextual bandit over fixed investigation prefixes. From a soft-computing perspective, the reward is an additive fuzzy measure whose rank-exponent weights act as an ordered weighted averaging (OWA)-type operator; one exponent sets its orness, i.e., how strongly the reward concentrates on the top of the judges' consensus.
We train an open-weight Qwen3-4B Triage Agent within a fixed multi-agent Distributed Denial of Service (DDoS) alert-triage workflow on DARPA 2000 LLDOS 1.0 and test it on an independent CSE-CIC-IDS2018 DDoS trace. Across three training seeds, the agent reaches a mean F1-score of 0.869, 5.9 and 9.7 points above uniform insight weights and DIVA, and a false-positive rate of 4.9% versus 16.4% for GPT-5-mini. Within the weighting family, false positives fall only at the highest-orness weighting evaluated. All claims are limited to DDoS alert triage; the code, task contexts, and ranked insight dataset will be released. - [58] arXiv:2611.17539 [pdf, html, other]
-
Title: Verifiable pXRD: Hill-climbing pXRD structure determination with LLM agentsComments: NeurIPS 2026 Workshop AI4Mat, spotlight (earlier version)Subjects: Artificial Intelligence (cs.AI)
Autonomous crystal-structure determination from powder X-ray diffraction (pXRD) remains a key bottleneck for self-driving materials laboratories. We formulate the task of pXRD structure determination as a verifiable agentic search problem: candidate structures can be assessed during inference through forward-simulated profile agreement and during evaluation through structural matching to a known reference. To benchmark progress, we introduce pXRD-Bench, a suite of 200 simulated single-phase problems spanning in-distribution structure determination, challenging and polymorphic structures, and controlled compositional and topological distribution shifts. Agent adaptation raises selected structural recovery from 27% to 69%, with further gains from repeated sampling and verifier-guided inference-time scaling. We extend this approach to experimental multiphase pXRD through a reflexive evolution framework that learns from self-generated synthetic problems. A replaceable laboratory specification defines the sample family and measurement conditions. An evolver agent uses the known generating structures and measurement parameters to diagnose solver failures and develop persistent guidance and executable tools, which are then provided to agents on measured patterns. This supplies verifiable training experience without requiring a collection of solved experimental measurements. On 15 measured patterns from solid-state synthesis, the evolved package improves average profile agreement and transfers an analogue-substitution procedure that recovers a phase whose structure is absent from the supplied reference library. These results demonstrate how learning from verifiable synthetic experience can improve agentic characterization of experimental samples whose structures are not known in advance.
- [59] arXiv:2611.17540 [pdf, html, other]
-
Title: Gusto: a toolkit for compatible finite element dynamical coresJemma Shipton, Thomas M. Bendall, Colin J. Cotter, David M. Acreman, Timothy C. Andrews, Jack Betteridge, Alex Brown, Thomas H. Gibson, Nell Hartney, David A. Ham, Joshua Hope-Collins, Jeremy A. Pike, Connor J. Ward, Dan WittSubjects: Numerical Analysis (math.NA)
Gusto is a tool for rapid prototyping of new algorithms based on compatible finite element methods for discretising the fluid equations in the dynamical core of weather and climate models. In this context, the development of new algorithms typically follows a path that involves testing on a hierarchy of simpler equation sets sharing important features of the mathematical structure of the equation set solved by the dynamical core. Gusto provides compatible finite element spatial discretisations of the hierarchy of equation sets, along with a range of timestepping schemes. The code is modular so that different spatial and temporal discretisations can be combined and the same schemes can be applied to any of the equation sets. This facilitates fair comparison of the schemes and makes the extension of any algorithmic developments to the more complex equation sets more straightforward. In this paper we present Gusto to the community by describing the algorithms implemented in Gusto and showcasing their performance through a variety of test problems commonly used in numerical weather prediction.
- [60] arXiv:2611.17541 [pdf, html, other]
-
Title: Quantum Algorithm with Quadratic Speedup for 2D Pattern MatchingSubjects: Data Structures and Algorithms (cs.DS)
We consider the problem of 2 dimensional pattern matching under Quantum Query model. Here, we are given a rectangular text $T[0..n_1-1][0..n_2-1]$ and a pattern $P[0..m_1 -1][0..m_2-1]$. We take $N = n_1n_2$ and $M = m_1m_2$ as the sizes of text and the pattern respectively (we assume $M < N$). Based upon the Quantum Advantage given by Grover's search (and subsequent Dürr-Høyer search), we create a solution which takes $\tilde{O}(\sqrt{N})$ query complexity to answer if the pattern $P$ exists in $T$.
Our solution uses deterministic sampling as a primary component. As noted by earlier work in the literature, deterministic sampling is inadequate and witness tables are essential components for 2D matching. In our settings, we assume that the witness table of the pattern is precomputed classically and is available as an oracle to the quantum algorithm. Our main novel component is how to compute and employ a geometric structure called ``skyline" without losing the quantum advantage. - [61] arXiv:2611.17548 [pdf, html, other]
-
Title: Mechanism-Level Chemical Reaction Simulation with Electron Bookkeeping TransformerSubjects: Computational Engineering, Finance, and Science (cs.CE); Biomolecules (q-bio.BM)
Modeling chemical reaction mechanisms is central to understanding and predicting chemical reactivity. We present Electron Bookkeeping Transformer (Bookkeeper), a data-driven model that simulates chemical reactions at mechanistic resolution by generating electron movement trajectories. Following the symbolic arrow-pushing formalism in chemistry, Bookkeeper explicitly models each electron as a token and predicts its movement sequentially, ensuring mass and charge conservation during both elementary step and reaction pathway predictions by construction. On reaction-size-balanced test sets, Bookkeeper improves top-1 elementary-step accuracy, top-1 pathway rank accuracy, and top-1 to top-5 product prediction accuracy over FlowER. On the entire test set, it achieves higher elementary-step top-1 accuracy but lower pathway accuracy than FlowER. Bookkeeper also samples elementary steps substantially faster, adapts to several unseen reaction types with minimal fine-tuning, and generates candidate chemical transformations for further screening. These results suggest that explicit electron bookkeeping is a useful framework for virtually screening chemical reactions across diverse reagents, validating computer-assisted synthesis planning predictions, and providing mechanistic interpretations of reactivity.
- [62] arXiv:2611.17549 [pdf, html, other]
-
Title: Talk to Act: A Synthetic Multi-Turn Dialogue Benchmark for Human Intent Extraction and Action Planning in RoboticsComments: 6 pages, 4 figures. This is an extended version of our paper accepted at the ICRA 2026 Workshop on Multimodal Embodied Interaction in Robots (MEI-2026), Vienna, AustriaSubjects: Robotics (cs.RO)
Multi-turn conversational human-robot interaction (HRI) continues to be a challenging problem in robotics, as many existing systems rely on rigid, command-driven interfaces that can be unintuitive for users. Recent advances in Large Language Models (LLMs) provide a promising direction, given their ability to process natural language, generate structured outputs, and support tool-based reasoning, all of which are relevant for interpreting user intent in robotic settings. However, progress in this area is constrained by the limited availability of datasets and benchmarks that capture the nuances of real-world human communication, such as multi-turn dialogue, ambiguity, and context-dependent requests.
In this work, we present a synthetically generated multi-turn dialogue dataset for HRI, annotated with action intents, object references, and task plans across a range of scenarios and user personas. Our primary focus is on evaluating the capability of small, open-weight language models for robotic tasks, particularly in settings where local deployment and efficiency are important. To this end, we frame intent extraction as a tool-calling problem where each predicted action corresponds to a structured function call with arguments grounded in the dialogue and benchmark models ranging from 270M to 8B parameters on their ability to perform intent extraction and action planning without task-specific fine-tuning.
Our results show that these models can achieve reasonable performance on the task even at relatively small scales and without any task-specific fine-tuning. This suggests that compact, locally deployable models already provide a useful baseline for conversational robotic systems. We expect that task-specific fine-tuning and more realistic training data will enhance performance further and make such models increasingly suitable for real-world robotic applications. - [63] arXiv:2611.17550 [pdf, html, other]
-
Title: Sub-Synchronous Control Interaction Risk of Large Data Center Loads: A Frequency-Scan Assessment with Cross-Model ValidationComments: 13 pages, 16 figures. Submitted for IEEE publicationsSubjects: Systems and Control (eess.SY)
Data centers are now one of the fastest-growing loads on the bulk power system, and most of their demand reaches the grid through power converters inside uninterruptible power supplies (UPS). Whether these loads can take part in sub-synchronous control interaction (SSCI), as wind and HVDC converters do, has not been studied extensively. We examine this question with impedance-based frequency scans at the point of common coupling (PCC). Using a detailed proprietary model (INS-ISU), we scan the motor, variable-frequency-drive (VFD), lighting, and UPS-fed information-technology (IT) branches of a representative facility separately. Only the UPS in double-conversion mode shows negative resistance; it covers the whole sub-synchronous band and continues up to about 170 Hz. The other branches, and the UPS in bypass mode, stay passive. We then repeat the scans on three public models from PNNL, TAMU, and NLR. None of them shows negative resistance in bypass mode, while three of the four models do in double-conversion mode. Lowering the load from rated to light load makes the negative resistance deeper, as expected for a constant-power load, so a lightly loaded facility in double-conversion mode is the worst case. Raising the UPS DC-link capacitance pushes the negative resistance into a narrow band around the fundamental. These findings give a basis for screening data centers for SSCI during interconnection studies.
- [64] arXiv:2611.17554 [pdf, html, other]
-
Title: Online Allocation with Differential PrivacyComments: Accepted by NeurIPS 2026Subjects: Cryptography and Security (cs.CR)
We study private online allocation problems, where allocation decisions must satisfy differential privacy (DP) to protect sensitive user information while optimizing performance under limited resources. We first formalize differential privacy notions suited to different privacy requirements in online allocation and establish the fundamental performance limits of any algorithm under these definitions. Based on these insights, we propose PPOA, a privacy-preserving meta algorithm for online allocation, and establish sufficient conditions under which PPOA preserves Joint DP (JDP) or Local DP (LDP) while achieving asymptotically near-optimal performance. Building on these conditions, PPOA can be instantiated into a variety of concrete algorithms. In particular, we present the specific designs, which include PPOA-DMD that preserves JDP and LDP and PPOA-FTRL that preserves JDP. We analyze their performance under both adversarial and stochastic settings and characterize the fundamental trade-offs between DP and allocation performance. Finally, we demonstrate the superior performance of the proposed algorithms via numerical experiments on AI model routing in battery-powered edge systems.
- [65] arXiv:2611.17555 [pdf, html, other]
-
Title: Red-Blue Hamiltonian Tomography via the Operator Lightcone WavemapComments: 15 pages, 2 figures, 3 tablesSubjects: Mathematical Software (cs.MS); Mathematical Physics (math-ph); Quantum Physics (quant-ph)
We introduce a model-free method for Hamiltonian tomography based on the operator spreading wavefront. Given two tensor network simulations of the same quantum lattice (one with the nominal Hamiltonian and one with one or more perturbed bonds), the signed cross-entropy of the non-identity Pauli weight field $\mathcal{L}(v,t) = -\ln(n_{\rm pert}/n_{\rm base})$ identifies each perturbed bond by its causal signature: a red-blue sign flip at the two endpoints, visible from the first wavefront crossing. Combining measurements from $p$ strategically placed observation centers via an ensemble log-likelihood sum, a single broken bond is identified with an 82-fold margin over the next candidate on a $7\times7$ lattice. Extending to realistic coupling degradation ($\delta J = -0.20$) and up to 8 simultaneous faults, the integrated log-ratio $\Lambda(v) = \ln(\mathcal{B}(v)/\mathcal{N}(v))$ achieves a per-bond detection rate above 90\% in a single seismic pass using only 6 simulations. An iterative re-scan strategy progressively resolves multi-bond cases as the signal clears with each recalibration pass. The method requires no model assumptions and no knowledge of the perturbation type or magnitude; the detection is deliberately blind.
- [66] arXiv:2611.17556 [pdf, html, other]
-
Title: A Synthetic Benchmark for Target-Centric Causal Discovery in Industrial ProcessesSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Causal discovery promises to find which recorded variables drive a process target such as product quality, yield or energy use. Process data with known causes are rare, and in existing benchmarks the difficulty of identifying the causes of a target is either fixed by the recorded data or set for the whole graph. We present a synthetic benchmark for target-centric causal discovery in industrial processes. Its generator is a two-level structural causal model inspired by batch production, assembled from components that each model one phenomenon of process data, such as a control loop or a delayed effect. The graph is built around the target: without sensor error, every non-cause can be distinguished from the direct causes by conditioning on them or by time order. Each component is set independently, so the difficulty can be varied one condition at a time. We evaluate eight methods, from association measures to effect estimation, with an error analysis by the role of the selected candidate; all degrade as the difficulties combine, each in a way consistent with its assumptions.
- [67] arXiv:2611.17560 [pdf, html, other]
-
Title: Generative Design with Real-time Physics Explainability: Towards Product Definition in the Age of AISubjects: Machine Learning (cs.LG); Computational Physics (physics.comp-ph)
In this article, we explore opportunities from using generative AI in mechanical engineering. On the one hand, generative methods based on invertible neural networks have been recently proposed for the solution of inverse design problems, allowing to infer design parameters from performance labels. On the other hand, generative AI models have shown that they can act as real time capable surrogates for physics based simulations. Here we suggest that these two streams should be combined in a unified framework, where INN are used to control the design process from the customer requirements perspective and the generative simulation of the systems physics adds a layer of explainability, why the proposed design fulfils the requirements. As generative design is prone to hallucinations, we propose uncertainty quantification to fend off physically infeasible requests. Finally, validation with classical tools is a requirement for generative design process. To demonstrate the capabilities of our novel generative design paradigm, we develop two interactive environments, one for airfoil design based on simulator data and one for the combustor design space exploration based on measured data.
- [68] arXiv:2611.17563 [pdf, html, other]
-
Title: Benchmarks Don't Talk Back: Learning Through Self-DisagreementComments: 34 pages, 5 figures, 15 tablesSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
A benchmark usually measures a language model alone with a question, and deployment rarely looks like that: answers get contested, and models give up correct ones under counter-reasoning that is itself wrong. We show that what fails there is not a shortfall of accuracy but a capability in its own right -- knowing when to hold a position and when to yield to a better one -- which we call dialectical competence. It stands on first-answer accuracy without following from it: a model can be made materially better at answering and no better at this, and training for accuracy alone can drift against it. Nor can it be bought at inference, by self-reflection or by sampling more. Taking a cue from the dialectical tradition in philosophy, in which productive opposition arises within a position rather than being imposed on it from outside, we hypothesize that this capability can be trained from opposition the model supplies itself. It states a position, generates the strongest case against that position -- its antithesis -- rules on whether the case stands, and is trained on that ruling, with no second model at training or at deployment. We call this learning through Self-Disagreement: reinforcement learning from verifiable rewards throughout, where the same check that scores an answer against its reference scores the ruling too, so no reward model, no preference data and no human annotation enter anywhere. The capability is trained rather than elicited: weak in the base model, it arrives with the reward, is not paid for out of first-answer accuracy, and holds against challengers the model never trained against, including ones stronger than itself. The opposition a model needs in order to improve, it turns out, need not come from outside it.
- [69] arXiv:2611.17564 [pdf, html, other]
-
Title: AI-Voyager: A Public-Interest System for Human-AI Civic Reasoning to Contest and Shape the Rules That Govern AISnehalkumar `Neil' S. Gaikwad, Dhruv Dixit, Bryen Alvarez, Pratik Patekar, Anika Sharma, Akshara Kolipaka, Paritosh PandeySubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
Governments are writing rules for artificial intelligence that decide what counts as an automated decision or a harm, often before the people they affect have a chance to question them. We introduce AI-Voyager, a public-interest system for human-AI civic reasoning, in which people and AI work together to understand, contest and shape the rules that govern AI. A resident can find their city's rules on AI in hiring or upload their country's AI strategy, read it as an affected worker would, open the passage behind any answer, and draft a revision. Its architecture turns 3,653 governance documents from city, state, federal and national governments into passages that keep their page and edition, and retrieves evidence through bounded agentic search that verifies each statement. On 360 new questions that require finding the relevant documents, it reached 83.0% macro recall, against 80.5% for the strongest baseline, 73.7% without iterative search and 99.1% on a benchmark with fixed documents. In a study with 24 participants new to two AI bills, most proposed changes asked how provisions would apply. Participants chose lenses for the kind of explanation they wanted, and citations signaled grounding but rarely led them to check the source. By binding AI assistance to the exact text of public documents, this work introduces a grammar of operations and a design space that inform how such systems are designed and evaluated, and how people take part in governing AI.
- [70] arXiv:2611.17565 [pdf, html, other]
-
Title: Performance-Portable Exact Gaussian Process Regression in Python via SYCL and oneMathComments: 12 pages, 6 figures, 2 tables. Accepted at the WACCPD 2026 workshop (SC26 Workshops), November 15-20, 2026, Chicago, Illinois, USASubjects: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG); Performance (cs.PF)
Modern multi-vendor GPU environments force a choice between portable code and native performance. GPRat is an open-source Python library for exact Gaussian process regression built on the HPX asynchronous many-task runtime, tiled algorithms, and CPU/CUDA BLAS backends. We contribute a UXL Foundation oneMath backend with a hybrid HPX/SYCL design, extending GPRat to AMD and Intel GPUs. The new SYCL backend comes within 3% of native CUDA on an NVIDIA A30 at the largest problem size and outperforms the GPflow and GPyTorch libraries by up to 1.75x and 2.17x on the A30 and up to 1.79x and 4.62x on an AMD Instinct MI210. On an Intel Arc B580, a consumer-class card with FP64 throughput below our CPUs, the unmodified SYCL backend trails GPRat's CPU backend. Overall, oneMath's dispatch to vendor-native BLAS lets a single SYCL codebase match CUDA on NVIDIA and outperform established Python GP frameworks on AMD, without vendor-specific code.
- [71] arXiv:2611.17566 [pdf, html, other]
-
Title: Grounding Object-to-Object Hazard Reasoning into 3D Risk Fields for Task-Conditioned SafetySubjects: Robotics (cs.RO)
Safety in robotics does not solely depend on geometric constraints, but also on semantics: the given task, environment properties, and interactions between objects in the scene. We present a framework based on vision-language models (VLMs) to infer task- and scene-dependent safety constraints for robot manipulation. We provide the VLM a structured safety tool based on an RGB-D observation and a natural-language task instruction that first decides whether the task should be rejected since its execution may result in a non-recoverable hazardous outcome. Otherwise, it identifies preventable object-pair hazards and specifies the spatial relations that must be avoided. Open-vocabulary segmentation and depth information are then used to ground the detected objects in 3D and construct continuous risk fields that represent the corresponding hazardous regions. These fields are integrated with an obstacle avoidance module that corrects nominal robot motions during execution while accounting for the risk field. We evaluate the framework on preventable hazards, tasks requiring rejection, and safe variants in simulation and the real-world. Project website: this https URL
- [72] arXiv:2611.17573 [pdf, other]
-
Title: CamSpokenNLI: A Pilot Resource for Cross-Variety Transfer Evaluation in Cameroon Pidgin English and CamfranglaisComments: 26 pages, 3 figures, 3 numbered tablesSubjects: Computation and Language (cs.CL); Sound (cs.SD)
Cameroon Pidgin English (a West African Pidgin English variety) and Camfranglais are widely used in Cameroon, yet to our knowledge, neither has a dedicated spoken Natural Language Inference (NLI) resource. We introduce CamSpokenNLI, a pilot dataset of 114 scenario-matched spoken NLI items in each variety (228 in total), recorded by one native speaker from Buea. Each item pairs premise audio and text with a hypothesis, an NLI label, and English glosses. The author assigned labels, so the dataset is an exploratory resource rather than a mature benchmark. The NLI experiments use the reference text; the audio is evaluated separately. XLM-RoBERTa baselines with five-fold item-level cross-validation reach 44.7% on Pidgin and 46.2% on Camfranglais (chance 33.3%). Still, neither is statistically distinguishable from hypothesis-only input, so hypothesis wording may account for much of the accuracy. Under disjoint-item cross-variety transfer, where a test item's counterpart is never seen in training, accuracy is 40.1% (Pidgin to Camfranglais) and 45.6% (Camfranglais to Pidgin). A counterpart-leak audit raises this to 60-64%, illustrating how counterpart leakage can overstate transfer. Zero-shot Whisper transcription of the recordings performs poorly: the best word error rates are 66.6% for Pidgin and 45.5% for Camfranglais, and with automatic language identification, Whisper Large-v3 labels 93 of 114 Pidgin recordings as Yoruba or Swahili. Together, the results suggest that such resources need disjoint-item evaluation and hypothesis-only controls, and that speech recognition must be adapted to these varieties before spoken NLI can be tested end to end.
- [73] arXiv:2611.17574 [pdf, html, other]
-
Title: Fault Tolerant CoresetsComments: 32 pages, 21 figures. NeurIPS 2026Subjects: Data Structures and Algorithms (cs.DS)
Coresets are small weighted summaries of large datasets that help quickly evaluate some loss objectives on the input dataset. We mainly focus on coresets for the $k$-means objective, where the cost of the input dataset with respect to a set of $k$ centers is the sum of squared distances from each point to its closest center. A $k$-means $\varepsilon$-coreset $\Omega$ satisfies $\mathrm{cost}(P,S)=(1\pm\varepsilon)\mathrm{cost}(\Omega,S)$ for every set $S$ of $k$ centers, where $P$ denotes the input dataset. A long line of work has established near optimal $\varepsilon$-coreset size bounds for $k$-means. A key assumption in these works is that the coreset remains intact after construction. The question we ask is the following: if an adversary corrupts at most $f$ weighted entries of an $\varepsilon$-coreset, can we recover an $O(\varepsilon)$-coreset via a decoding algorithm that is given only the corrupted coreset. This models situations where entries of the coreset can be corrupted during storage or transmission, and recovery must proceed without access to the original dataset. Our main result shows that increasing the usual coreset size by an additive $O(fk/\varepsilon)$ yields a coreset that can be decoded into an $O(\varepsilon)$-coreset after any $f$ adversarial corruptions. This bound is achieved by sensitivity sampling, a widely used method for constructing coresets. Our result is optimal in the sense that any fault tolerant coreset, whose decoding algorithm only keeps a subset of the points must also have size $\Omega(fk/\varepsilon)$. Our experiments demonstrate recovery under adversarial corruptions.
- [74] arXiv:2611.17575 [pdf, html, other]
-
Title: Natural Swing-Up of an Underpowered Double PendulumSubjects: Robotics (cs.RO); Systems and Control (eess.SY)
Robots with weak motors must ride their natural dynamics rather than fight them. We swing up a direct-drive planar two-link robot, whose motors deliver only about 6\,\% of its shoulder gravity torque, along its nonlinear normal modes, two-dimensional invariant manifolds consisting of periodic orbits. On a non-strict mode, injecting energy necessarily moves the state off the manifold, so pumping and stabilization compete for the same bounded torque. We split the input torque into a power-neutral steering component and a scalar pumping input, the only component that changes the energy. The allocation then becomes a linear program in two variables: maximize pumping subject to the torque limits and one control-Lyapunov constraint. The constraint uses the model-based rate of the error, so no stabilizing feedback gain is needed. Experiments in hardware show swing-ups along both modes and compare the allocation with summing-and-clipping and with prioritized stabilization.
- [75] arXiv:2611.17576 [pdf, html, other]
-
Title: Beyond a Single Sign: BiTE for Branch-Preserving Binary Threshold ExpansionComments: 8 pages, 5 figuresSubjects: Machine Learning (cs.LG)
Binary neural networks (BNNs) reduce model storage and replace many arithmetic operations with efficient bitwise computations, but the single-sign activation limits their representational capacity. Each preactivation direction is compared with one threshold and reduced to one of two values, discarding variation within each sign region. We propose \textbf{BiTE} (Branch-Preserving Binary Threshold Expansion), which evaluates each feature direction against $K$ ordered thresholds and retains the resulting binary responses as separate inputs to the following learned operator. BiTE can distinguish up to $K+1$ ordered activation states while allowing the following operator to learn a separate transformation for each threshold response. It also enables a controllable trade-off between the number of independently constructed feature directions and the number of ordered responses extracted from each direction. On BinaryViT, BiTE improves mean final-epoch Top-1 accuracy from $74.60\pm0.30\%$ to $75.83\pm0.15\%$ on CIFAR-10 and from $42.80\pm0.46\%$ to $44.07\pm0.32\%$ on CIFAR-100, with nearly unchanged trainable parameter counts. A matched uniform-collapse control reaches $75.33\pm0.15\%$ on CIFAR-10, supporting the value of retaining the threshold responses separately relative to the evaluated scalar collapse. Binary VGG-16 experiments further show a favorable observed parameter--accuracy trade-off, including a configuration that achieves $0.64$ percentage points higher mean accuracy with $21.8\%$ fewer parameters than conventional width scaling. Overall, these results support branch-preserving threshold expansion as an effective way to improve binary representations and the observed parameter--accuracy trade-off.
- [76] arXiv:2611.17577 [pdf, html, other]
-
Title: Fair and Efficient Distributed EV Charging Coordination Using Threshold Cryptography and Machine LearningComments: 16 pagesSubjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Systems and Control (eess.SY)
Distributed coordination of electric vehicle (EV) charging with the alternating direction method of multipliers (ADMM) keeps EV constraints private but assumes that every participant follows the algorithm. This assumption fails in practice, where a biased operator or a man-in-the-middle (MITM) attacker can send inconsistent signals to different EVs, and compromised EVs can deviate from their prescribed updates. We first prove that attacks by such arbitrarily malicious, i.e., Byzantine, participants do not prevent convergence and therefore remain undetected, and derive the most damaging attack in closed form. We then propose a framework that combines a threshold-signed broadcast with upload recomputation, which detects every deviation without false alarms at linear instead of quadratic communication cost. To keep the optimization running, each affected EV recovers locally with a certified, fleet-aware machine learning price forecast, and we prove that the optimization remains convergent with a recovery cost bounded by the shape of the forecast error. Simulation results with real-world load, price, and charging data show that a biased operator raises the annual cost of disadvantaged EVs by 132.7% and lowers that of advantaged EVs by 49.3% without changing the aggregate load, whereas the defense limits all cost changes to 2.6%. An external MITM attacker raises the cost of its victims by 57.6%, the peak load by 5.8%, and the grid cost by 15.8%, whereas the defense keeps them within 1.1%, 0.2%, and 0.3% of benign operation. With up to 70% of the EVs attacked, the recovery converges in at least 97.8% of the runs.
- [77] arXiv:2611.17581 [pdf, html, other]
-
Title: Beyond Source Code: Evaluating Natural Language Specifications for LLM-Based C-to-Rust TranslationSubjects: Software Engineering (cs.SE)
Large language models (LLMs) have shown promising results for C-to-Rust translation, yet existing approaches primarily rely on source code as input, leaving natural-language specifications of program behavior underexplored as an alternative or complementary representation. We hypothesize that direct code translation encourages LLMs to preserve C-specific implementation choices, producing Rust that may be functionally correct but unnecessarily verbose, reliant on unsafe constructs, and less idiomatic. We investigate this hypothesis by comparing three translation settings: code-only, specification-only, and code-and-specification. Using 300 C programs from the IBM CodeNet dataset, we evaluate GPT-5.2 and GPT-4o-mini on functional correctness, maintainability, complexity, idiomaticity, unsafe construct usage, and cost. Code-only translation achieves the highest test pass rate (98.6%) but often mirrors C's implementation patterns. In contrast, specification-only translation produces the most idiomatic and compact Rust, reducing unsafe blocks by approximately 6x and Clippy style warnings by 2.4x. However, using specifications without validation yields test pass rates as low as 86% with our best performing model. Adding an iterative test-based validation and refinement step for specifications raises this to a 95.3% test pass rate. Overall, specification-based translation consistently improves code quality relative to code-only translation, while code-and-specification offers the best balance between correctness (98.3% test pass rate) and maintainability. Our study systematically compares these translation settings, highlighting the trade-offs among specification quality, translation correctness, code quality, and computational cost.
- [78] arXiv:2611.17583 [pdf, html, other]
-
Title: When Does Less Information Pay? An Exact Finite Atlas of Representation, Payoff, and CostComments: 40 pages, 5 figures; supplementary material included. Code, data and Russian translation: this https URLSubjects: Artificial Intelligence (cs.AI)
A representation that retains one bit about a four-state world can solve a task perfectly, while another that retains 1.5 bits on average reaches only 75% accuracy. The difference lies in which distinctions each representation preserves. We examine this relationship in a uniform four-state world by enumerating all admissible representations and binary action rules. Constructions are compared both by replacing one component while holding the others fixed and by choosing the best action rule separately for each module. We consider three tasks, their mixtures, and four separate hypothetical cost tariffs.
In some comparisons, the representation retaining less world information has the higher payoff; in others, the more informative representation does. There are also pairs with different information amounts but equal payoff--that is, equal task accuracy. If a representation preserves all existing distinctions and adds new ones, its best zero-cost payoff cannot decrease. Lower-information advantages therefore depend on which distinctions are retained. Equal counts of opposing advantages under representation replacement follow from coding symmetry. In broader comparisons, the shares depend on how repeated realizations are counted. A different task set or resource price can change which member of a fixed pair performs better.
These results are exact, complete censuses of the specified finite sets of constructions and comparisons. The atlas provides a reproducible worked model for examining how action constraints, counting rules, and cost assumptions affect comparisons. It illustrates established distinctions between information quantity and decision value rather than proposing a new general law of their relation. $I(W;R)$ measures retained world information, and payoff measures task accuracy. - [79] arXiv:2611.17585 [pdf, html, other]
-
Title: Rethinking Visual Reasoning in Text-to-Image Reward ModelingSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Reward models are a core component of modern text-to-image systems, serving as the objective for inference-time selection and RL-based post-training. A key prerequisite is visual understanding: to judge prompt adherence and ground aesthetic preferences, a reward model must bind attributes, resolve relations, and count reliably. We find that widely-used reward models perform poorly on targeted visual reasoning evaluations. Even when starting from strong vision-language backbones, preference fine-tuning erodes these capabilities, and continual-learning baselines fail to mitigate a steep tradeoff between visual reasoning and preference accuracy. We propose a data-centric intervention that injects visual reasoning supervision into preference learning by flipping the pairing axis. To complement conventional human preference data (paired images under a shared prompt), we convert existing visual reasoning datasets into preference pairs that share the same image but differ in prompt, and train on the mixture. This yields a Pareto improvement in both visual understanding and preference prediction. Across inference-time best-of-n selection and RL post-training, the resulting reward model improves text-to-image generation on challenging compositional prompts (GenEval) by an average of 4.0% and 5.5% respectively, outperforming HPSv3 which was trained on 6x more preference pairs.
- [80] arXiv:2611.17586 [pdf, html, other]
-
Title: Where to Predict in Retinal OCT? Anatomy-Guided Target Selection for I-JEPAComments: 18 pages. Accepted as a poster at the GenAI for Health Workshop at NeurIPS 2026 (non-archival)Subjects: Computer Vision and Pattern Recognition (cs.CV)
Self-supervised learning aims to learn meaningful representations from unlabeled data, potentially allowing models to capture subtle anatomical variations across patients and reduce the amount of labeled data needed. Image-based Joint-Embedding Predictive Architecture (I-JEPA) learns by predicting target-region representations from visible context rather than reconstructing pixels. However, its uniform target placement does not account for the anatomical structure of retinal OCT, where retinal tissue occupies only part of each B-scan. We investigate anatomy-guided target selection using image intensity and predicted retinal segmentation while retaining the I-JEPA architecture and latent prediction objective. We compare anatomy-guided rectangular targets, anatomy-shaped targets and coverage-constrained placement with uniform masking, and evaluate the learned representations using linear probing for glaucoma classification on FairVision. We found that rectangular targets guided by image intensity achieved the highest linear-probe AUC in our comparisons. We show that anatomically meaningful target selection is promising for retinal OCT representation learning, while greater anatomical specificity did not consistently improve performance.
- [81] arXiv:2611.17587 [pdf, html, other]
-
Title: An Algebraic Perspective on Operational SemanticsSubjects: Programming Languages (cs.PL)
Languages with effects are typically given two independently specified semantic accounts: a denotational semantics based on (free) models, and an operational semantics defined by effect-specific reduction rules. This separation leads to two independently constructed semantic structures whose relationship must then be established a posteriori. We present a unified construction in which both denotational interpretation and operational reduction are constructed from an algebraic theory specifying the effects.
We work with call-by-push-value extended by the operations of an algebraic theory. We prove equational canonicity and give a tree-based operational semantics that reduces every closed computation to a tree of operations. When the free model monad is a monadic container, we derive from it a configuration-based small-step semantics. Configurations are elements of the free model. Every step preserves the denotation. Under a condition we call affinity, every reduction terminates and at ground types computes the denotation. We instantiate the framework with state, errors, writer, and weighted and pre-convex probabilistic choice. For errors, writer and weighted choice we recover the familiar rules. For state we get a machine that runs the program from every initial state at once. - [82] arXiv:2611.17589 [pdf, html, other]
-
Title: Small Open-Weight Language Models for Crowd Decision-Making: Capabilities and Limits of Output-Level Fusion and Linguistic AggregationDavid Herrera-Poyatos, Cristina Zuheros, Carlos Peláez-González, Andrés Herrera-Poyatos, Rosana Montes, Francisco HerreraSubjects: Computation and Language (cs.CL)
Small open-weight language models can run crowd decision-making (CDM) on an institution's own infrastructure, which keeps the checkpoint, the inference settings and the opinions under its control. Their judgements vary across repeated queries; whether repeating and aggregating them, often through the 2-tuple linguistic representation, settles the decision rather than the labels is untested. Five open-weight models of 8-36B total parameters on one consumer GPU and a hosted reference classify 474 TripAdvisor reviews of four London restaurants ten times each; three open-weight models repeat the protocol on 600 reviews of twelve Madrid restaurants. Continuous evaluations preserve resolution: snapping each review leaves 5 distinct values against 21-65, and snapping at every level at most two collective values. Sensitivity to the participants persists once run-to-run variability is reduced: in London, redrawing the reviewers keeps the winner on 47-81% of replicas against 86-100% for redrawing the runs. With simultaneous intervals over every pair of alternatives under crowd resampling, no system on either corpus names a single winner: in London every system keeps all four restaurants as candidates, and in Madrid the candidate sets hold five to eight of the twelve. Extra runs and output-level fusion stabilize the winner but do not remove that sensitivity, and the fusion does not beat its best member. The three larger open-weight models name the hosted reference's London winner and at least match its per-review correlation with the reviewers' stars, so small open-weight models can support local CDM, reported as candidate sets with margins rather than forced rankings.
- [83] arXiv:2611.17596 [pdf, html, other]
-
Title: Tractable Importance Sampling for Monte Carlo Privacy AccountingSubjects: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
Evaluating tight Differential Privacy (DP) parameters for DP matrix factorization with privacy amplification by sampling requires analyzing high-dimensional, correlated Gaussian mixtures. Because integrating these mixture distributions to find the exact privacy failure probability ($\delta$) is analytically intractable, these mechanisms are typically analyzed using Monte Carlo accounting. However, standard sampling suffers from prohibitive variance when estimating small $\delta$ associated with strict privacy guarantees, which in turn requires a large sample complexity, making Monte Carlo accounting inefficient. We demonstrate that Monte Carlo accounting can be made much more efficient via importance sampling. We give a shifted importance sampling proposal distribution parameterized by a stretch factor $\alpha > 1$. We prove a closed-form bound $B(\alpha) < 1$ on the support of the importance-weighted estimator that holds identically for both directions of the hockey-stick divergence, show that $B$ is log-convex so that the optimal $\alpha$ is a one-dimensional root-finding problem, and show that the resulting estimator seamlessly plugs into the KL-Hoeffding verification protocol used by past work, with a sample complexity reduced by exactly a factor of $1/B(\alpha^*)$.
- [84] arXiv:2611.17598 [pdf, html, other]
-
Title: X-Nav Technical Report: Real-World Long-Horizon Indoor-to-Outdoor Navigation on a Humanoid RobotJiaxin Huang, Chengxing Lin, Xiaoyu Luo, Sirui Chen, Jingyu Guo, Aolin Li, Kairui Chen, Rui Zhen, Cheng Wen, Kerui Gu, Shunyao Wang, Zhihui Peng, Shaoli HuangSubjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
Long-horizon navigation from indoors to outdoors remains an open problem for humanoid robots: one instruction covers rooms, doors and roads, and the camera moves with the gait. Worken, AGIBOT's measure of productive robot work, depends on robots that reach their workplaces autonomously. We present X-Nav, to our knowledge the first system to take a bipedal humanoid through a language-instructed long-horizon mission from indoors to outdoors. X-Nav covers a pipeline from data to deployment. Its data engine collects navigation data in open-source scenes, in digital twins of deployment sites and on real robots. Its navigation model, MELIOR, takes only RGB images from the robot's built-in camera and comes in two versions: MELIOR-A predicts actions, and MELIOR-N predicts a navigation goal. Its agentic workflow, X-Navarch, turns a brief human instruction into sub-tasks and deploys MELIOR on the robot with explicit target navigation and obstacle avoidance. On the Lingxi X2, X-Nav completes a mission of about a hundred metres from an office lobby to a store in a neighbouring building; LiDAR and odometry are used for target approach and low-level control. The mission is completed with both versions of MELIOR: the robot follows a change of destination and walks through a door from the lobby to the road, and with X-Navarch it also passes static and dynamic obstacles on the route. During deployment, MELIOR is adapted with data collected in a digital twin of the site; the adaptation raises its closed-loop success rate in the twin, and the adapted model is then deployed on the robot at the real site. In addition, MELIOR achieves competitive results on the R2R-CE and RxR-CE benchmarks.
- [85] arXiv:2611.17600 [pdf, html, other]
-
Title: LightTAR: Light and Fast Target-Aware Representations for Multimodal Tabular LearningComments: 21 pages, 3 figures. Code: this https URLSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
Multimodal tabular learning involves classification and regression on tables combining structured numerical and categorical features with unstructured text and images. These unstructured fields are commonly represented by pretrained encoder embeddings reduced with principal component analysis (PCA). This reduction prioritizes input variance but may overlook information relevant to predicting the target. We introduce LightTAR, a simple, inexpensive method that adapts dimensionality reduction to the prediction task. LightTAR augments the standard PCA representation with a few target-aware features extracted jointly from the final and an intermediate encoder layer using regularized partial least squares (PLS). We evaluate LightTAR with 21 tabular models, including tree ensembles, neural networks, and tabular foundation models. LightTAR improves average predictive performance for all models and, averaged across models, improves prediction on almost every MulTaBench dataset. These gains come with a small increase in total runtime relative to PCA representations. LightTAR integrates directly into existing encoding pipelines, improving tabular learners' predictive performance without modifying the encoder or the learner.
- [86] arXiv:2611.17602 [pdf, html, other]
-
Title: LinearKnot: Linear-Time Approximation of the Pseudoknot Partition FunctionSubjects: Data Structures and Algorithms (cs.DS)
Pseudoknots, where two or more helices cross, are central to many functional RNAs, including frameshifting signals, telomerase RNA, riboswitches and viral genomes. With arbitrary pseudoknots, finding the minimum free energy structure is NP-hard, so most RNA secondary structure prediction tools only consider pseudoknot-free structures. Dirks and Pierce (DP03) pioneered the first pseudoknot partition function, which underlies NUPACK's pseudoknot model, but it takes $O(n^5)$ time and $O(n^4)$ memory, making it practical only for very short sequences. Linear-time approximate dynamic programming, as in LinearFold and LinearPartition, does not transfer easily to DP03: a partial pseudoknot leaves a gap filled by its crossing partner, whose complementary sequence may still lie ahead of the scan, so it cannot be ranked by its inside weight. We present LinearKnot, a linear-time approximation of DP03 in three passes: (a) LinearPartition gives pseudoknot-free weights; (b) a second pass builds tables that estimate each partial stem's crossing partner; and (c) a third scans a left-to-right reformulation of DP03 that reserves the 5' nucleotides in the pseudoknot helix for subsequent pairing, keeping the states with the largest estimated contribution. Without pruning, this reformulation computes DP03's pseudoknot recursion exactly, so the only approximation is which states the beam keeps. End-to-end, time grows linearly ($n^{1.01}$) and memory marginally above linear ($n^{1.15}$) from 1,000 to 9,392 nt. LinearKnot folds the 9,392 nt Rous sarcoma virus genome in 2.4 hours and 90.5 GB, where exact DP03 would need an estimated 40,000 years and 54 million GB, and matches exact DP03 base-pair probabilities on 418 PseudoBase RNAs with a median RMSD of $9.2\times10^{-7}$. Our code is available at this https URL.
- [87] arXiv:2611.17603 [pdf, html, other]
-
Title: Offline KL-Regularized Reinforcement Learning with Stochastic TransitionsSubjects: Machine Learning (cs.LG)
KL-regularized reinforcement learning is the objective behind RL fine-tuning of language models, and its data are often collected by the reference policy itself. However, beyond bandits and deterministic transitions, little is known about how much offline data it needs. In this work, we study finite discounted Markov decision processes with unknown stochastic transitions, $K$ state-action pairs and effective horizon $H$. For data that cover the regularized optimum with concentrability $C\geq 4$, every estimator needs $\Omega(KCH^3/(\epsilon\max\{\epsilon,\beta\}))$ samples on some instances when $\epsilon$ is small, where $\beta$ is the temperature. Under a fixed design, pessimistic soft value iteration with a variance-aware penalty attains this rate with high probability, up to logarithmic factors, when $\beta\geq 1.4H$. When the data come from the reference policy itself, we show that it covers the optimum up to a power law whose exponent grows as $H/\beta$; in the same temperature range, this power law alone gives the estimator a sample complexity polynomial in $1/\epsilon$, with no coverage assumption. For $\beta\geq H$ the power can also increase the sample size needed by a factor polynomial in $1/\epsilon$. Finally, the plug-in estimator, which is minimax optimal with a generative model, can fail on reference data at small temperatures when many actions are rarely seen, while the pessimistic estimator does not.
- [88] arXiv:2611.17605 [pdf, html, other]
-
Title: Choosing Quantized Code LLMs by File Size and Energy per Solved TaskComments: 9 pages, 2 figures, 3 tables. Submitted to GREENS 2027 (11th International Workshop on Green and Sustainable Software, co-located with ICSE 2027). Reproducibility package: this https URLSubjects: Software Engineering (cs.SE); Machine Learning (cs.LG)
A developer running a code assistant on a laptop downloads one model file, and local inference moves the energy of every request onto that machine. A common rule says a larger model at lower precision tends to beat a smaller one at higher precision. We ran 24 GGUF files of Qwen2.5-Coder at four sizes on a fanless MacBook Air M2 and priced each in energy per solved task on 542 EvalPlus tasks. The result is a selection guide indexed by Ollama's listed file size, with an evidence label on each pick. At 1 and 2 GB it picks the q4_K_M files of the 1.5B and 3B models, both on the correctness and energy frontier under every accounting. At 3 GB as listed, the larger-model rule picks 7B-q2_K, the only measured 7B file that fits, and there it does not pay. That file brings no significant gain in solves over the 2 GB pick, 3B-q4_K_M, and the 90% CI limits any 3B-q4_K_M deficit to 2.4 points. Yet 7B-q2_K needs 2.03 times the energy per solved task on CPU, GPU and memory rails, or 1.66 to 1.81 times on whole-system power. On HumanEval+, five alternating session pairs give a ratio of 1.925 against 1.956 in the main grid. At 8 GB, 7B-q6_K buys a significant gain over the 2 GB pick at a higher energy price. From 1.5B up, q4_K_M trails the highest precision measured at its size by less than 3 points (90% CI). We release per-task results, power logs and the analysis script.
- [89] arXiv:2611.17608 [pdf, html, other]
-
Title: Neural Simulation-Based Inference for Computational Cognitive Models: A TutorialSubjects: Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
Computational cognitive modeling is an increasingly essential methodology to elucidate the mechanisms underlying human and animal cognition. By fitting models to empirical data, researchers can perform parameter estimation, model comparison, and model-based neural analyses to assess competing theories, quantify individual differences, and investigate neural representations. Typically, fitting these models relies on likelihood-dependent methods (e.g., Maximum likelihood estimation or hierarchical Bayesian methods). However, for many reasonable cognitive models, computing the likelihood is computationally intractable, in which case these methods become inadequate. This substantially limits the range of theories that researchers can investigate. Simulation-based inference with neural networks (neural SBI) has emerged as a powerful toolkit to overcome this hurdle. By training deep neural networks on datasets generated through model simulations, neural SBI learns the relation between observable data and cognitive models. This approach enables all the applications of standard modeling while bypassing the likelihood computation, without constraints on empirical data size. Despite its successful adoption across diverse scientific fields, neural SBI remains underutilized in cognitive modeling, partly due to a lack of accessible, domain-tailored guidance. This tutorial addresses this gap by providing a clear workflow for applying neural SBI to cognitive models. We also provide practical use cases and recommendations for reducing the computational and simulation burden, facilitating efficient implementation in real-world research settings. By lowering the technical barriers to neural SBI, this tutorial aims to empower cognitive scientists to investigate richer theory spaces and ultimately propel scientific discovery.
- [90] arXiv:2611.17610 [pdf, html, other]
-
Title: More Agents, More Samples, or More Thinking? Untangling Coordination, Ensembling, and Test-Time Compute in Multi-Agent LLMsComments: 43 pages, 11 figures, 13 tables. PACE Reviewer Checklist, evaluation utilities (pace CLI), and reproduction data at this https URLSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
Evaluations of multi-agent language models often compare communicating architectures against a single greedy pass. This practice confounds coordination with sample multiplicity and inference compute. We synthesize nine primary empirical evaluations comprising 516 unique multi-agent configurations organized along a resource vector R with six dimensions and a baseline ladder with four categories. Against baselines with a single pass (Category A), multi-agent accuracy differences are negative overall and near zero when omitting the largest contributing study. When evaluated against baselines matched on sample count (Category B) or resource budget (Category C), multi-agent systems average -0.042 in accuracy, and multilevel random effects modeling on the log odds scale remains negative across sensitivity checks. Multi-agent accuracy equals or falls below the baseline in 74% of these comparable contrasts, and study means are nonpositive in almost all evaluated studies. Across studies reporting both baselines, moving from a greedy pass to a sampled baseline shifts accuracy differences by -0.038 on average. Because communication adds no information when agents share identical evidence, positive returns require partitioned evidence, external verification, or error diversity across models. We provide a reviewer checklist and accounting utilities.
- [91] arXiv:2611.17612 [pdf, html, other]
-
Title: Battery Energy Storage Sizing in PV Ramp-Rate Smoothing using Constraint-Aware Reinforcement Learning-Based ControllerSubjects: Systems and Control (eess.SY)
This paper presents a constraint-aware reinforcement learning (CARL) controller for battery energy storage system (BESS) sizing and photovoltaic (PV) ramp-rate smoothing. Rapid PV power fluctuations caused by cloud transitions can introduce operational and stability challenges in power systems. To mitigate these impacts, grid codes impose ramp-rate limits (RRL) on PV power injected to the grid. BESSs can support such grid codes by absorbing or supplying power during rapid variations. However, the extent to which BESSs can reduce PV-induced variability depends on (i) the BESS power and energy capacities; and (ii) the adopted controller. Thus, BESS sizing and controller design are coupled. Conventional moving-average control can increase battery usage due to filter-induced lag, while ramp-rate control is reactive and does not account for the long-term impact of actions on battery state of charge and future ramp events. Optimization-based methods can reduce BESS requirements by anticipating future PV variations, but their performance depends on accurate PV forecasts. To address these limitations, the CARL controller learns a causal state-feedback policy from historical PV data without requiring forecasts during deployment. Ramp-rate, converter-power, and state-of-charge constraints are incorporated into a causal feasibility interval that guarantees feasible actions by definition. A full-horizon input-convex neural network (ICNN)-based optimization is developed as a noncausal benchmark for the minimum achievable BESS requirements. Evaluation using three real-world PV datasets shows that, at a 10\%/min ramp-rate limit, CARL reduces peak BESS power by 8.7--10.4\% and energy capacity by 47.5--62.6\% relative to conventional ramp-rate control while maintaining the grid-codes compliance.
- [92] arXiv:2611.17616 [pdf, html, other]
-
Title: LLM-Guided Cluster Lifecycle Management: Deciding When and How to Evolve Behavioral ClustersComments: 16 pages, 8 figures, 9 tablesSubjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Behavioral clusters learned over user or item interactions drift as behavior evolves: some fragment into sub-behaviors, some converge, some disappear, and new patterns emerge. Two paradigms dominate the maintenance of evolving clusterings. Threshold triggers fire a pre-specified action when a hardcoded statistic crosses a cutoff; refit-and-track (MONIC-style) re-clusters every window and infers transitions from how the new clusters align with the previous window's. Both hardwire the decision into a threshold or a re-clustering rule. We study an alternative in which a large language model (LLM) makes the lifecycle decision in place. Given a threshold-free, whole-population view of the current cluster set (statistical health, labels, member texts, residual mass), the LLM decides per cluster whether and how to act (split, merge, retire, create, or no-action) without re-clustering. Because the useful signal channel depends on how informative the labels are, we fix a richness-conditioned policy in advance: statistics when label distinctiveness is low, fused statistics+semantics when it is high. On three public datasets spanning label richness (YooChoose 0.17, MOOC 0.50, News-Category 0.70), the policy-selected LLM mode obtains the highest F1 on all three, with far fewer interventions and stronger restraint than tuned threshold heuristics and a MONIC steelman (paired Wilcoxon p <= 0.003, bootstrap 95% CIs excluding zero). The gain is concentrated more in precision than recall. The useful modality shifts from statistics to semantics/fusion as richness rises. Semantics hurt on YooChoose, fusion is roughly neutral on MOOC, and the semantic channel is required on News. One base prompt template and one fixed policy serve all three datasets, whereas the tuned baseline collapses when moved between them. We treat this shift as a practitioner diagnostic observed on three datasets, not a validated law.
- [93] arXiv:2611.17618 [pdf, html, other]
-
Title: Lean-PEP: Performance-Estimation Certificates to Formally Verified Algorithmic BoundsComments: 26 pages, 2 figures, 8 tablesSubjects: Logic in Computer Science (cs.LO)
Performance-estimation certificates describe worst-case bounds for optimization algorithms. Turning a certificate into a formal theorem requires exact verification of its algebra and a connection between its coordinates, constraints and objective and the intended function-level problem. Lean-PEP provides a reusable interface for this task. It checks a submitted source and exact certificate against a supported canonical finite model, and a parameterized analytic theorem transfers the resulting finite bound to every admissible algorithm trajectory. Generated computation proofs discharge the obligations that depend on each certificate. An additional capture entry checks the numerical record and reconstruction policy. The implemented two-step proximal-gradient consumer verifies a new input with normalized bound 62500/499999, and a fixed eight-configuration batch yields seven distinct payloads with accepted certificates and executed Lean proofs. PG/GD cases and one-step attainment witnesses establish the corresponding algorithmic conclusions, and a fresh installation reproduces consumption of the supplied input. These results show how new certificate data can reuse a proved analytic interface to obtain function-level bounds within the supported models.
- [94] arXiv:2611.17619 [pdf, html, other]
-
Title: TouchTwin: Scaling Human Touch Data via Real-to-Sim-to-RealHuiwen Han, Nuo Chen, Ziyao Zeng, Dayou Li, Hao Wang, Amir Hossein Rahmati, Xiaoning Qian, Bangya Liu, Zhiwen FanSubjects: Robotics (cs.RO)
Human demonstrations are becoming the main source of data for dexterous manipulation, and real-to-sim-to-real pipelines multiply each demonstration in simulation. These pipelines rebuild the interaction from vision, which recovers how the hand moves. Which fingers touch, when, and how hard remain unobserved, because a hand pose shows where each finger stops on the surface rather than how far it is driven to close. This gap is a force gap: on held-out recordings, a hand policy that imitates the observed pose grips with 0.31x the recorded force, and a policy trained on physics replays of vision-tracked poses grips with 1.55x. TouchTwin closes the gap with measured human touch. Each selected TouchScale recording, captured with head and wrist RGB-D cameras and an 880-taxel glove, is rebuilt in MuJoCo. Coding agents write the objects as physics-tested programs, and a per-subject tactile hand replays the interaction with the recorded touch setting the contact and the closing force of each finger. In simulation, a policy trained on touch-grounded replays grips at 0.99x the recorded force and succeeds in 20/20 trials on held-out recordings and 35/36 unseen perturbations; a contact predictor trained on them misses 13% of real contact frames, against 32% for vision-based replays. Each replay also seeds perturbed demonstrations with simulated contact and force labels, rendered as realistic video in new views and scenes.
- [95] arXiv:2611.17620 [pdf, html, other]
-
Title: Safety Side Effects of Preventative Persona Steering: An Empirical AuditComments: 5 pages, 1 figure (6 charts), 1 table, accepted at NeurIPS 2026 Bialign WorkshopSubjects: Artificial Intelligence (cs.AI)
High-level behaviors such as sycophancy and hallucination correspond to directions in a model's activation space (Chen et al., 2025). These directions, known as persona vectors, can suppress undesirable traits either at inference time, subtracting the vector during generation, or through preventative steering, adding the vector during fine-tuning, so the optimizer never needs to encode the trait in weights, then removing the vector at evaluation. Chen et al. (2025) recommend preventative steering because it preserves Massive Multitask Language Understanding (MMLU) accuracy, but validations of this recommendation have measured capability alone. We primarily evaluate results on preventative steering of sycophancy and hallucination vectors on three models, measuring three safety benchmarks alongside trait expression, coherence, and MMLU. Our findings show that preventative steering still changes refusal behavior, but by an amount that depends on the trait and does not establish that preventative persona steering is safety-neutral, and steering methods should be evaluated on several safety benchmarks.
- [96] arXiv:2611.17621 [pdf, html, other]
-
Title: AscendGYM: A Cost-Aware Kernel Optimization Benchmark for Real-World Workloads on Ascend NPUsAI Training Platform Team: Bingwei Zhang, Jing Li, Amir Abdullin, Sihang Chen, Xin Chen, Liangjun Feng, Yingmeng Gao, Tao Guo, Junjie Huang, Dongfang Li, Zhanglian Lin, Jiajun Liu, Zekun Liu, Zhengxuan Lu, Xiaodong Luo, Jian Meng, Ivan Nosov, Linyuan Qiu, Weijian Sun, Wanrui Yan, Binfan Zheng, Zhicheng Zhou, Baotian Hu, Min Zhang, Haizhou Li, Zhiquan LuoComments: 20 pages, 12 figures, 2 tables. Technical Report. Code available at this https URLSubjects: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
Kernel optimization is essential for efficient large language model (LLM) training and inference on Ascend NPUs. Existing Ascend kernel benchmarks primarily focus on kernel generation, with less emphasis on the ability of agents to optimize existing implementations under input configurations drawn from real workloads. To address this requirement, we introduce AscendGYM, a benchmark whose initial suite comprises 48 kernels across three difficulty levels. Each kernel is paired with test cases whose data types and tensor shapes are collected from real-world training and inference tasks. Consequently, performance improvements on AscendGYM map directly to the corresponding workloads. In terms of usability, two evaluation modes are supported in AscendGYM: model evaluation and model-harness evaluation. The kernel optimization capabilities of both trained models and custom agents are allowed to be measured. It provides a complete execution environment that closely reproduces how Ascend kernels are invoked in practical pipelines. For model evaluation, it additionally includes a built-in harness equipped with search orchestration and experience accumulation. Finally, we propose a new metric, the Speedup-Token Score (STS), which quantifies performance gains relative to token cost. By jointly considering speedup and cost, STS enables the assessment of practical kernel optimization capability of a model or harness under an acceptable budget. AscendGYM is available at this https URL.
- [97] arXiv:2611.17622 [pdf, html, other]
-
Title: Auditing Chatbots for Incursion into Professional Legal PracticeNel Escher, Megan Elliott, Isabelle Qian, Yana Sharifullina, Svenja Guhr, Deja Davis, Deirdre MulliganSubjects: Computers and Society (cs.CY)
The provision of legal advice has long been in the exclusive jurisdiction of licensed attorneys-a boundary that unauthorized practice of law (UPL) doctrine maintains to protect the public from incompetent counsel. We argue that this boundary is being redrawn, not by courts or bar authorities, but by frontier AI companies. While early chat models were trained to avoid dispensing legal advice, modern models have been reconfigured to reduce refusals and prioritize helpfulness. We ask whether chatbots increasingly do tasks that require a law degree and bar admission, and whether this is harmful in high-stakes legal domains. We develop UPLBench, a benchmark derived from court decisions finding UPL in high-stakes legal domains (e.g., immigration, family law). We then audit 23 chatbot models dating from 2023 to present, evaluating whether chatbots refuse or provide substantive legal advice, and whether they refer users to lawyers or legal aid. Unlike prior work that aims to establish the legal capabilities of chatbots (e.g., bar exam performance), our approach converts legal doctrine into a testable benchmark for evaluating chatbot adherence to law. We find modern chatbots increasingly perform tasks courts reserve for lawyers and decreasingly recommend that users consult lawyers. Subject matter expert attorneys reviewing our transcripts identified interaction patterns, such as failure to elicit relevant facts, that would be considered inadequate counsel. If acted upon, such advice could lead to deportations, loss of homes, and family separations-the very harms that UPL rules are designed to protect against. We weigh paths forward to protect the public, including compelling models to comply with existing law. We call upon the legal profession to respond to this incursion into its domain of expertise as there are insufficient protections and remedies for users who rely on chatbot legal advice.
- [98] arXiv:2611.17626 [pdf, html, other]
-
Title: Joint World Action Models through Differentiable Trajectory OptimizationComments: 29 pages, 6 figures, 14 tablesSubjects: Machine Learning (cs.LG)
Joint world action models (WAMs) generate future actions and states together, modeling world-action trajectories rather than actions alone. However, trajectory imitation does not ensure that these trajectories are task-effective or consistent with dynamics. We present \textbf{world action trajectory optimization (WATO)}, which optimizes generated actions and future states jointly for control. WATO uses both as references in a dynamics-constrained trajectory optimization problem and differentiates the task loss through the optimized trajectory to train the generator. This makes generated trajectories task-aware and dynamics-consistent while using trajectory optimization only during training. At deployment, WATO directly generates candidate trajectories and ranks them with a world model, without iterative trajectory optimization. Across $7$ visual control environments, WATO improves average success from $79.9\%$ for the imitation joint WAM to $91.7\%$, outperforming strong model-based, goal-conditioned RL, and generative control baselines. Compared with cross-entropy method (CEM), WATO is on average $11.5\times$ faster and uses $600.5\times$ fewer world model evaluations at deployment.
- [99] arXiv:2611.17629 [pdf, html, other]
-
Title: Training-Free Acceleration of Multi-View 3D Transformers via Regime-Aware Attention ApproximationSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Multi-view 3D transformers such as VGGT reconstruct a scene from a set of unposed images in a single forward pass, predicting camera poses, depth and 3D points, but the cost of their global attention grows quadratically with the num- ber of images. Accelerators borrowed from large language models assume that each token attends to only a few oth- ers; on these models they lose accuracy or gain little speed. We study the global attention of three multi-view transform- ers on seven datasets and find that every layer plays one of three roles: it averages over the whole scene, matches points across many images, or anchors on the first image, whose coordinates the outputs use. Building on this anal- ysis, we propose Regime-AWare ATTention Approximation (RAWATTA), a training-free method that runs attention in 8- bit precision and keeps 16-bit precision only in the layers where a worst-case calibration check requires it. Its fast set- ting, RAWATTA-Fast, also approximates attention in a few averaging and anchoring layers, while matching layers and the first image keep all their keys. On a 12 GB GPU at 50 im- ages, RAWATTA runs VGGT faster than the released code, mostly through our CUDA engine, and loses at most about one point of pose accuracy (AUC@3) on any dataset; RAWATTA-Fast is slightly faster and less accurate. Both fit 200 images, where the released code runs out of memory at 100.
- [100] arXiv:2611.17630 [pdf, html, other]
-
Title: A mesh-free deep energy method for phase-field modeling of thermally driven brittle fractureHan Zhang, Mehrisadat Makki Alamdari, Babak Shahbodagh, Mohammad Vahab, Cosmin Anitescu, Timon Rabczuk, Elena AtroshchenkoComments: 48 pages, 18 figures, 15 tablesSubjects: Machine Learning (cs.LG); Materials Science (cond-mat.mtrl-sci)
Phase-field simulation of thermally driven fracture requires resolving localized crack bands whose paths are unknown in advance and thermal boundary layers whose thickness evolves during the transient. Neural energy-based computations of thermal fracture have so far started from defects placed in advance. This paper proposes a mesh-free deep energy method based on multiresolution neural representations that predicts spontaneous crack nucleation, crack spacing, and selective crack propagation and arrest under transient thermal loading. Two networks, each with its own multiresolution encoding, represent the displacement and phase fields and the transient temperature, and the mechanical energy and the incremental functional of the backward Euler heat equation are estimated on integration points redrawn at every iteration. Restricting irreversibility to the crack set, carrying the optimizer state across load steps and decaying the learning rates within each step help prevent spurious damage accumulation. On two notched plates the initiation peak loads agree closely with an independent finite element reference. On a quenched glass strip under an analytical temperature field the crack path changes from straight to oscillatory and then to branched, with an oscillation threshold close to the linear stability prediction. Under a prescribed boundary temperature a quenched ceramic slab forms a crack array of the type observed in experiment, with a spacing close to that of a published benchmark. On the slab and a quenched annulus the arrays are denser than those of a finite element computation on the same material field. No crack paths, nucleation sites or crack-adapted meshes are prescribed.
- [101] arXiv:2611.17631 [pdf, html, other]
-
Title: On the Utility of Pre-Trained Visual Representations for Policy Learning of Contact-Rich Insertion TasksSubjects: Robotics (cs.RO)
Visuomotor imitation learning has become a predominant paradigm for real-world robotic manipulation, yet it remains data-inefficient and brittle to task-irrelevant visual shifts. Pre-trained visual representations promise to ease this burden by distilling priors from internet-scale data, but the ``right'' representation for robotics remains unresolved: existing studies span broad but sparse task distributions and reach conclusions that vary and even conflict, pointing to strong task and setting dependence. We examine this question by bringing a magnifying glass to a common contact-rich manipulation task: insertion. We compare representations from widely used vision foundation models against an explicit contact-centric representation on two closely related insertion tasks, perturb each along controlled axes of occlusion and physical tolerance, and test whether conclusions transfer to the real world. Task dependence proves to be more subtle than previous, broadly scoped evaluations revealed: the best representation differs even between our near-identical tasks, and small changes shift trends in unpredictable ways. Despite this variability, results across simulation and real-world experiments demonstrate the distinct value of spatially-grounded, and in particular, contact-centric information for contact-rich policy learning, while underscoring how difficult it is to estimate such cues reliably under realistic visual complexity and sensor noise. We hope these findings motivate further study of representations in contact-rich manipulation, along with advances in contact perception and its integration into visuomotor policy learning.
- [102] arXiv:2611.17632 [pdf, html, other]
-
Title: Generated Evaluation Suites as Diversity-Driven Search for Real Agent RegressionsComments: 9 pages. Accepted at the NeurIPS 2026 EvoRobust WorkshopSubjects: Artificial Intelligence (cs.AI)
Generating an evaluation suite from an agent's specification is diversity-driven search: a generator fills a descriptor space (requirements $\times$ personas $\times$ scenario kinds), and the suite it emits is an archive, then scored like a benchmark by its mean pass rate, which throws away the search's product. We inject 7 controlled defects into a $\tau^2$-bench agent across two domains. Coverage yields a ceiling: if a defect's traced footprint holds $b$ scenarios that pass at baseline, it can itself contribute at most $b/N$ to a suite of $N$. For 4 of 5 traceable defects that ceiling lies at or below the detection threshold $\theta = 0.10$, so no within-footprint effect can move the suite past it; in both domains the tool-removal defect destroys every covered scenario it can (exactly its ceiling) while the suite-level pass rate moves the other way. Defects split into two regimes: trace-aligned ones, where per-cell scoring recovers the tool-removal defect in full and reveals a second as a significant improvement, and trace-orthogonal ones, whose footprint is behavioural: forced escalation is detected in retail at a traced coverage of exactly zero. Diversity along the generator's own axes helps only where it aligns: at fixed suite size a happy-path-only suite detects nothing, while a kind-diverse one detects the orchestration defect in 5 of 20 draws. Evaluation archives should be scored per cell and should publish coverage as a first-class property.
- [103] arXiv:2611.17638 [pdf, other]
-
Title: AI-guided robotic in situ fabrication of curvature- and tissue-adaptive wireless bioelectronicsSubjects: Robotics (cs.RO)
Personalized bioelectronic devices for wound therapy require not only conformal fabrication on complex body surfaces but also functional adaptation of their electrical characteristics to individual wound geometries. Here, we present an AI-guided robotic in situ printing platform for fabricating wireless wound electrotherapy devices directly on transparent wound dressings. The system integrates a 3D vision camera, a six-degree-of-freedom robotic arm, direct-ink-writing, and geometry-aware RF design optimization. AI-based wound segmentation is used to determine wound dimensions and local surface curvature, while the reconstructed three-dimensional geometry is incorporated into the electromagnetic design of the receiving coil and interdigitated electrodes. In contrast to conventional approaches that use surface geometry only for trajectory generation, the proposed method accounts for both anatomical curvature and tissue dielectric loading to compensate for geometry-induced shifts in resonant behavior. A printable silver ink was characterized in terms of rheology, dimensional reproducibility, electrical resistance, and electromechanical stability, and was used to fabricate multilayer wireless devices directly on curved polyurethane dressings. Wireless power transfer, impedance characteristics, and electric-field delivery were evaluated experimentally and numerically. Finally, robotic in situ fabrication was demonstrated on an anthropomorphic tissue-equivalent forearm phantom, where the measured dielectric properties of the phantom were incorporated into the RF optimization, enabling resonant operation near the target frequency of 25 MHz. This work establishes a perception-to-design-to-fabrication framework for geometry-adaptive wireless bioelectronics and provides a technological foundation for future physical-AI-enabled manufacturing of personalized on-body electrotherapy devices.
- [104] arXiv:2611.17639 [pdf, html, other]
-
Title: What LLM Agents See and What Users SaySubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Before an LLM agent cancels a reservation or deletes a folder for a user, what stops it from acting on a false assumption: what its tools show, or what the user says? In this work, we study this gap with presupposition theory from linguistics: an agent may act only if each claim behind the act (e.g., one reservation fits the name) is seen to hold, answered by the user, or disclosed while the act can be undone. A simple bound shows that when lookups cannot tell one record from two, any agent acts without asking first as often with two records as with one. We then vary whether a search lists, counts, or omits a second match, and whether the user states a claim that a record refutes (e.g., "It is empty"). Under a role-only prompt, a count of a second record leaves most of 36 earlier agents acting on the one they see, whereas listing both stops most. When the user says a record such as a folder is empty, however, a count of its one item stops agents about as often as a listing, and stating the claim mostly cuts acting when it is false. The newest Claude and GPT models always open a counted second record, then never act. When we add a same-name, same-zip customer to a retail benchmark, a lookup returning that customer first led to wrong-account writes in about a quarter of conversations, and one that checks a street the user knows kept completion near the one-customer level. Safe writes therefore need lookups that list or check competing records, and evaluations that vary what tools show.
- [105] arXiv:2611.17644 [pdf, html, other]
-
Title: CoTTAGE: An Explainable Judge for Text-to-Audio GenerationYu-Xiang Lin, Yun-Man Hsu, Bo-Ru Lu, Kuan-Po Huang, Wenze Ren, Ming-To Chuang, Chun-Yi Kuan, Jiawei Du, Byeonggeun Kim, Qingming Tang, Chieh-Chi Kao, Ankur Gandhe, Chao Wang, Hung-yi LeeSubjects: Sound (cs.SD); Artificial Intelligence (cs.AI)
Rapid progress in text-to-audio (TTA) generation demands reliable and explainable automatic evaluation. However, existing metrics fail to simultaneously deliver high correlation with human judgments and meaningful, audio-grounded explanations, forcing reliance on labor-intensive human evaluation that bottlenecks model development iteration. To address this, we propose CoTTAGE, a reasoning-based evaluator for text relevance (REL) and audio quality (OVL). Following a reason-then-predict paradigm, CoTTAGE first generates an audio-grounded rationale to diagnose errors before predicting scores. Supported by over 10K newly collected human OVL annotations, CoTTAGE is first trained via supervised fine-tuning and then refined with reinforcement learning using a Thurstone-based preference objective that converts sampled score distributions into pairwise preference probabilities and rewards agreement with rankings derived from human judgments. Across six benchmarks spanning REL and OVL, CoTTAGE achieves state-of-the-art correlation with human judgments while producing human-preferred explanations to accelerate TTA development.
- [106] arXiv:2611.17646 [pdf, html, other]
-
Title: Binarization as an Attack Surface in Weightless Neural NetworksLucas Senos Coutinho, Matheus Vargas, Pedro Teixeira, Heudson Tosta Mirandola, Ori Rottenstreich, Priscila Machado Vieira Lima, Leandro Santiago, Daniel Sadoc MenascheComments: This paper appears at the Trustworthy Machine Learning Workshop (ARTMAN 2026) co-located with the IEEE Annual Computer Security Applications Conference (ACSAC 2026)Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Weightless Neural Networks (WNNs) are lightweight classifiers that binarize their inputs and learn by storing binary patterns in lookup memories, which makes them attractive for resource-constrained tasks such as intrusion detection. Because a small perturbation that crosses no binarization threshold leaves the binary input unchanged, WNNs may appear naturally robust to adversarial examples; a perturbation that does cross a threshold, however, can still change the prediction. We study this tension in WiSARD, the canonical RAM-based WNN, and in two variants that replace its memories with Bloom filters, Bloom WiSARD and BTHOWeN. A continuous perturbation reaches these classifiers only by crossing an encoding threshold, changing at least one binary tuple, and redirecting a memory query; we formalize this mechanism with a threshold-crossing bound and a response-margin certificate for standard Boolean WiSARD. Controlled MNIST experiments compare linear and data-driven thermometer encodings under model-agnostic sensitivity probes, a score-based Simulated Annealing attack, and surrogate-transfer attacks. Changing only the encoding while keeping each model otherwise fixed shows that encoding granularity can trade a small clean-accuracy loss for a large decrease in attack success. On the Bot-IoT, CICIDS, and UNSW-NB15 intrusion-detection datasets, adversarial perturbations lead WNN detectors to classify many attacks as normal traffic and to raise false alarms, even when their aggregate accuracy remains high.
- [107] arXiv:2611.17650 [pdf, html, other]
-
Title: Classifier-Guided Latent Diffusion for Small-Molecule GenerationSubjects: Machine Learning (cs.LG)
Diffusion models have emerged as a promising class of generative models for molecular design because of their ability to learn complex chemical distributions and support conditional or guided generation. In this work, we investigate a latent diffusion pipeline for property-guided small-molecule generation. The proposed framework represents canonicalized SMILES strings using a frozen ChemBERTa encoder, trains a Transformer-based decoder to reconstruct SELFIES strings from molecular latent embeddings, and learns an unconditional denoising diffusion model in the ChemBERTa latent space. To enable property-directed generation, a differentiable latent property predictor is used to guide the diffusion sampling process. Generated SELFIES strings are converted to SMILES, validated using RDKit, and evaluated using molecular validity, uniqueness, novelty, and predicted property scores. We study guidance toward two drug-relevant binary endpoints: blood--brain barrier permeability and inhibition of human $\beta$-secretase 1. The results suggest that latent diffusion over pretrained molecular embeddings can generate chemically valid and novel molecules while enabling controllable steering toward desired molecular properties.
- [108] arXiv:2611.17658 [pdf, html, other]
-
Title: Imitation Policies Learn a Proprioceptive Anchor: A Predictive Law for Perception Noise and Its Closed Loop CostComments: 18 pages, 4 figuresSubjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Imitation policies for robot manipulation are often trained on noisy object positions, and they learn to discount the noisy reading. We show how much they discount it, and why. Because demonstrators steer toward the object, the robot's own state carries information about where the object is. A trained policy combines this proprioceptive anchor with the reading and weights the reading by $\alpha = V/(V + s^2)$, where $s^2$ is the reading noise and $V$ the anchor's error variance. $V$ belongs to the learner and is far smaller than a linear model of the demonstrations implies, which is why classical errors-in-variables (EIV) regression overestimates $\alpha$. Measured once, $V$ predicted ten held-out sensitivities across two robomimic tasks, two learners and four new noise levels, with mean absolute error 0.025 against 0.37 for classical EIV, and a parameter-free bias test rules out the simpler explanation that the policies just overestimate the noise. The discount does not carry over to deployment, where the policy produces its own state: closed-loop sensitivity to a persistent reading offset exceeded $\alpha$ in all 15 policies that passed the registered gates, although the stricter registered prediction, that all would follow the offset fully, held for 8 of them. Post hoc, lower $\alpha$ went with lower success, steeply for diffusion, and a registered prospective test confirmed this for diffusion on Can at two new noise levels: the measured $\alpha$ predicted success before any rollout, within 0.03 (95% intervals $\pm 0.08$), although training noise predicts it equally well. Every test was registered before its data, including the classical benchmark that failed, and all other analyses are labelled post hoc; all results are in simulation with synthetic Gaussian noise.
- [109] arXiv:2611.17659 [pdf, other]
-
Title: INC-SRv6: In-Network Route Exploration and Fast Failover Using Segment Routing Headers in Dynamic UAV NetworksComments: 22 pages, 8 figures. Submitted to Electronics (MDPI)Subjects: Networking and Internet Architecture (cs.NI); Distributed, Parallel, and Cluster Computing (cs.DC)
Unmanned aerial vehicle (UAV) networks require a network provisioning mechanism that guarantees delivery and latency quality of service (QoS) under dynamic conditions in next-generation airborne mesh networks by continuously learning and repairing multi-hop paths under rapid link turnover without flooding the control plane. This paper presents INC-SRv6, an in-network computing routing mechanism that repurposes the IPv6 Segment Routing Header (SRH) and its Segments Left (SL) pointer into a self-learning, fast-failover distributed control plane. Rather than treating the SRH as a static source-committed vector, INC-SRv6 uses the active SL pointer to partition the segment list in flight: slots ahead of SL provide capacity for hop-by-hop next-hop recording, while the suffix behind SL serves as an in-band multi-destination reachability vector. Through this dual-zone header discipline, transit routers extract distance-weighted reachability updates from arriving suffixes, gate admission by single-pass confidence with bounded binary probing, advance or reverse SL deterministically, and restore pre-transmission SRH state for local failover upon link loss. Simulations across 200-node 3D UAV networks show that under high-speed flight and repeated 20% relay outages, INC-SRv6 improves packet delivery over reactive routing (92.12% versus 89.12%) while cutting 95th-percentile tail delay from 121.26 to 8.84 ms and control packets by 35.8%, with a 4-byte compact uSID stride model reducing SRv6 control bytes by over 52%.
- [110] arXiv:2611.17660 [pdf, html, other]
-
Title: EViDS: An Environment--Vehicle Dynamics Integrated Simulation Platform for Intelligent Chassis DevelopmentSubjects: Robotics (cs.RO)
The development of intelligent chassis increasingly requires the integration of environmental and perception information with high-fidelity vehicle dynamics to support environment-aware intelligent chassis control. However, existing autonomous-driving simulators and professional vehicle-dynamics tools are generally developed for different purposes, making it difficult to integrate environmental information, vehicle dynamics, and chassis algorithms. This paper proposes EViDS, an Environment-Vehicle Dynamics Integrated Simulation Platform for Intelligent Chassis Development. EViDS integrates CARLA-based environment and sensor simulation with CarSim-based high-fidelity vehicle dynamics. Reference-pose-based state alignment and multi-rate interaction maintain spatial and temporal consistency, while ray-casting-based tire-road interaction transfers environmental information to the vehicle-dynamics model. Scenario configuration, synchronized data acquisition, and algorithm interfaces further support intelligent chassis development. The platform is evaluated through environment-dynamics interaction tests with vehicle-dynamics consistency comparisons, together with intelligent chassis applications. Results show that EViDS preserves vehicle-dynamics characteristics, reliably transfers environmental information, and supports online closed-loop intelligent chassis control, providing a unified simulation basis for environment-aware intelligent chassis development and validation.
- [111] arXiv:2611.17661 [pdf, html, other]
-
Title: Feasibility of Pareto-Improving Damping Control of Forced Oscillations: The Active-Power CaseSubjects: Systems and Control (eess.SY)
Damping control of a forced oscillation may reduce amplitudes at certain locations while increasing them elsewhere. This paper introduces Pareto improvement as a criterion for assessing such control: at least one nodal oscillation amplitude decreases, while none increases across all other nodes. The analysis considers active-power forcing sources and active-power control injections in connected swing networks with positive inertia and damping. Necessary conditions for Pareto improvement feasibility are derived from the normalized nodal responses. It is proven that for every network with a single oscillation source and a single actuator, at least one positive forcing frequency is shown to preclude Pareto improvement, regardless of the amplitude and phase of the control input. A three-node counterexample shows that this limitation does not persist with multiple actuators. PSS/E simulations of a 6-bus benchmark system and a modified IEEE 39-bus system validated the analysis. These findings suggest the value of assessing control effects across the network and further investigating multi-actuator damping control.
- [112] arXiv:2611.17662 [pdf, html, other]
-
Title: Learn then Bet: Learning Interpretable Rules for Reliable Agent InterventionSubjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Autonomous agents now run for hundreds of steps with little or no human supervision. In practice, practitioners use runtime monitors such as turn limits and keyword filters to decide when a run should be stopped, upgraded, or assisted. However, these heuristics may stop runs that would otherwise succeed or allow failing runs to continue. To address this limitation, we introduce LeBet (Learn then Bet), a statistical learning framework that derives its learning objective from sufficient conditions for valid and powerful intervention. In particular, LeBet learns an interpretable OR-of-AND monitor and accumulates evidence through a fixed betting statistic, intervening when the evidence reaches a prescribed threshold. Under finite within-class exchangeability and leave-one-out stability, we derive finite-sample guarantees for false-positive control (validity) and detection power conditional on the training data, using the same trajectories for learning and calibration without assuming independent and identically distributed trajectories. To address the intractable combinatorial optimization problem in our learning formulation, we use stochastic local search that leverages the monitor's Boolean structure to propose rule edits aimed at improving empirical separation. On SWE-bench Lite with Ministral 3 8B, the learned monitor saves $34.46\%$ of input-plus-output tokens and increases output precision from $18.33\%$ to $21.59\%$. The empirical false-positive rate is $13.64\%$. Our results show how to formalize the learning problem over interpretable rule sets with statistical conditions and establish training-conditional guarantees for reliable intervention based on partial histories.
- [113] arXiv:2611.17669 [pdf, html, other]
-
Title: SZZBench: Controlled Git Histories with Explicit Oracles for Diagnosing SZZ ImplementationsComments: 5 pages, 1 figure, 2 tablesSubjects: Software Engineering (cs.SE)
The SZZ algorithm traces a bug-fixing commit back to the commits that introduced the bug. Its labels underlie defect-prediction data sets and many mining software repositories studies. SZZ implementations are mostly evaluated against oracles mined from real projects. Such oracles are realistic, but each case mixes the history factors that make an implementation succeed or fail, and an aggregate score cannot show why. We present SZZBench, a tool that shows why: it generates ordinary Git repositories from declarative scenarios with explicit oracles, scores the predictions of any SZZ implementation under declared evaluation policies, and reads each result against 26 naive baselines, such as always answering the second commit. SZZBench also checks the declarations of new scenarios, such as the tier, against the results of the naive baselines. Currently, SZZBench ships 279 scenarios in 17 categories and adapters for four SZZ tools. Running these tools in nine configurations without changing their code, we find that PyDriller passes 161 scenarios and a rule that always answers the second commit passes 160, yet only 97 scenarios are passed by both: the two succeed on different scenarios, which the number of passed scenarios hides and the per-scenario results reveal. SZZBench is available at this https URL.
- [114] arXiv:2611.17671 [pdf, html, other]
-
Title: Provable Cross-Modal Transfer of Hierarchical Concept Structure in CLIP EmbeddingsComments: NeurIPS 2026Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Contrastive vision-language models such as CLIP are trained to align images and text, not to represent taxonomies. Yet empirical studies show that CLIP embeddings exhibit hierarchical structure aligned with the hierarchy of semantic concepts. A theoretical characterization of why such hierarchy emerges from a non-hierarchical objective remains open. In this paper, we formalize this phenomenon as Hierarchical Angular Separation (HAS), a geometric property whereby each non-leaf concept embedding is more similar to its descendants than to off-branch concepts. Empirically, we show that HAS holds approximately in raw images and randomly initialized CLIP image embeddings, while in text embeddings it holds approximately only after contrastive learning. To theoretically explain this asymmetry, we analyze an unconstrained features model in which the image embeddings are fixed and only the text embeddings are trained. At the global minimizer, the text-side HAS violation percentage at each node is at most the image-side HAS violation and near-violation percentage. The key mechanism is the hierarchical co-occurrence structure of image-text pairs, which biases text embeddings toward root-centered image prototypes. For a linear contrastive loss, we derive a closed-form solution that makes this mechanism explicit. We extend the analysis to the InfoNCE loss and establish analogous guarantees under sufficient regularization. Synthetic experiments validate the theory and show that HAS is transferred even beyond the regimes covered by our analysis.
- [115] arXiv:2611.17672 [pdf, html, other]
-
Title: Meta Build Systems as Logic ProgramsComments: 60 pages, 4 figures; includes supplementary materialSubjects: Programming Languages (cs.PL); Software Engineering (cs.SE)
Meta build systems such as CMake, Meson and Bazel derive command graphs from project descriptions, user choices and machine facts. Imperative scripts, search paths and special cases make routine decisions interact poorly: search order can choose compatible packages without an explicit project choice, packagers must patch author flag preferences, and cross-builds need annotations or nested build trees for tools run during the build.
Our central insight is to express configuration as relational inference: objects follow from sources, flags and libraries from dependencies, and language standards from transitive requirements. We introduce buildlog, a meta build system whose stratified logical core isolates build derivation from impure environmental probes. It rejects ambiguity in toolchain or package selection and explains derived facts through provenance. Author constraints and user preferences use distinct, statically checked input relations. Libraries can ship consumer rules, including Qt's moc/rcc rules, using the same facts for installed and built code. Using a compiled program in a build rule lets buildlog infer its build-machine configuration.
We evaluate buildlog on eleven case studies of ten projects, validating configurations through execution and artifact checks while reporting remaining flag differences and platform limits. In Halide, generators link an LLVM-based compiler and run during the build to emit WebAssembly or Android code; where CMake requires nested builds, buildlog derives build-machine and target-platform configurations in one graph. A supporting formal account establishes termination, within-round probe-order independence, and agreement between fresh and validated cached runs under explicit finiteness and probe obligations. The results show that confining impurity to fact discovery yields a small, uniform and explainable configuration semantics. - [116] arXiv:2611.17674 [pdf, html, other]
-
Title: Pattern-Based Matrix Analysis Under PermutationsComments: Accepted at NeurIPS 2026 (The Fortieth Annual Conference on Neural Information Processing Systems)Subjects: Machine Learning (cs.LG)
Low-rank approximation captures one important notion of matrix simplicity: a matrix is simple if it is well described by a low-dimensional linear model. This paper studies a complementary notion. We consider matrices that become structurally simple only after permuting rows and columns, so that their mass concentrates in a prescribed geometric pattern such as a band, block-diagonal, arrowhead, or related form. We introduce a pattern distance that measures how close a matrix is to a given pattern after optimal reordering, giving a common framework for structural problems such as latent ordering, community detection, and bandwidth minimization. The problem is NP-hard in general; we derive an alternating algorithm based on Linear Assignment subproblems. We further introduce a pattern-plus-low-rank decomposition that combines permutation-structured fitting with a low-rank residual, whose permutation step is pattern fitting on the residual; its gains across pattern families persist at a matched parameter budget and are substantially reduced under a density-matched random-mask control. We evaluate the framework on synthetic recovery tasks, real correlation and graph adjacency matrices, and transformer attention heads, showing that a single universal algorithm recovers qualitatively different forms of matrix structure.
- [117] arXiv:2611.17680 [pdf, html, other]
-
Title: Aligned, Conflicting, or Unspecified? A Diagnostic Benchmark for Multimodal Time Series ForecastingComments: Accepted at IEEE AIoT 2026 WorkshopSubjects: Artificial Intelligence (cs.AI)
Artificial Intelligence of Things (AIoT) systems increasingly combine sensed time series with external text such as expert reports and news, yet how reliably models use text that agrees or conflicts with the sensed history remains unclear. We introduce the first real-world diagnostic benchmark organized by text-history relationship: 1,762 text-window pairs across sensor telemetry (Weather), web-service traffic (Wikipedia), and market data streams (Finance), each labeled Aligned, Conflicting, or Unspecified by whether the text's direction agrees with the historical trend, opposes it, or is not stated. Across ten pretrained models evaluated zero-shot, we investigate text selection and combination under text-history relationships (RQ1) and error diagnosis and horizon-scaled revision control (RQ2). For RQ1, we compare two types of text on paired Weather windows: one directly forecasts temperature changes, while the other describes related weather conditions, such as winds and fronts. Under Aligned, the first type yields the lowest error for most models. Under Conflicting, it increases most models' error above the no-text baseline; we therefore add only its stated magnitude and timing to the second type. Compared with either type alone, this combination lowers mean absolute scaled error (MASE) for six models and raises directional accuracy (DA) for seven. For RQ2, an error decomposition attributes the error increase to oversized revisions, and a training-free rule that limits them outperforms forecasts both with and without text for the four models with this trade-off, lowering MASE by 4.49%-21.98% and raising DA by 2.99%-9.71%. Our dataset is available at this https URL.
- [118] arXiv:2611.17681 [pdf, html, other]
-
Title: OCTAVE: A Capability-Resolved Benchmark for End-to-End HTML5 Game GenerationKehang Li, Xinlan Xu, Yulin Jin, Heng Cao, Xuchen Li, Houxian Ren, Zihao Ji, Qiang Luo, Yadong Mao, Qilin He, Jian Zhang, Jian Zhao, Zihao WangSubjects: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
Complete game generation requires agents to coordinate rules, interactions, states, and presentation over a long development process. A single aggregate score obscures which capabilities are implemented and which requirements remain unmet. We present OCTAVE, a benchmark for complete HTML5 game generation organized around eight Capability Dimensions, with 180 tasks and 1,800 Evaluation Requirements drawn from real player requests, commercial-mechanics reconstruction, and human expert design. It covers PC Web, Mobile Web, 2D, and 3D. Requirements distinguish Completeness, assessing publicly specified core implementation, from Richness, assessing hidden, deeper system relationships. Static source-code judgments provide requirement-level diagnosis across capabilities and task conditions. Evaluation of 11 model-harness combinations yields 33.2% Completeness and 32.5% Richness; 19.9% of generation records are static gate failures or missing artifacts. Excluding same-model self-evaluation, even the best combination remains below 50% on both public and hidden requirements. Richness is not harder than Completeness under every task condition. Systems perform relatively better on Core Rules and Numerical Design, while AI/NPC and Feedback and Game Feel show more prominent gaps. Requirement difficulty offers coarse stratification but is only weakly negatively correlated with pass rates. A limited human cross-check finds that requirements judged satisfied by humans but failed by the automated evaluator are the main direction of disagreement. OCTAVE extends evaluation beyond aggregate scores to diagnosis localized by requirement, capability, and task condition.
- [119] arXiv:2611.17684 [pdf, html, other]
-
Title: Representational Geometry as a Diagnostic for Adversarial RobustnessSubjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
Adversarial training is a common technique for making neural networks more robust to small, intentionally crafted input perturbations, and robustness is commonly measured through accuracy under attack. By exploring inter-network representations across different adversarial training objectives, we can analyze whether multiple networks organize their latent representations similarly. We study Multilayer Perceptrons trained on MNIST under three adversarial training regimes and compare the resulting population training dynamics and geometry to those of a vanilla network. We use M-PHATE to embed data samples across training epochs and examine how the geometry of the data changes during training under adversarial objectives. On penultimate-layer representations we measure displacement between clean and perturbed points, distributional distance through maximum mean discrepancy, and nearest-neighbor recovery. Our visualizations and quantitative analyses show that adversarial training leads to structural changes in how networks organize internal data representations, and robust accuracy and representational stability can rank the same defenses in opposite ways. These findings support a discussion of re-evaluating how practitioners approach the auditing of model robustness by treating latent representations and their training trajectories as diagnostic artifacts across sets of networks.
- [120] arXiv:2611.17686 [pdf, html, other]
-
Title: What Do Multilingual Self-Supervised Speech Encoders Know About Sentence Meaning?Comments: Accepted to EACL 2027Subjects: Computation and Language (cs.CL)
Self-supervised speech models (S3Ms) have been scaled up to thousands of languages and millions of hours of audio, yet little is known about how these models structure semantic knowledge across languages. In this work, we probe sentence-level semantics in massively multilingual S3Ms across 76 languages, using linear maps from frozen speech representations into a frozen text embedding space, evaluated by cross-modal retrieval. Sentence meaning is linearly accessible but highly uneven across languages and resource tiers. Cross-lingual probes can match language-specific probes on languages never seen in probe training, but per-model monolingual decodability does not predict transfer. Within the OmniASR wav2vec family, scaling from 300M--7B disproportionately benefits high-resource languages, and ASR tuning sharply improves semantic decodability while making representations more language-specific. However, semantic content is entangled with acoustics, inflating apparent performance on single-speaker corpora.
- [121] arXiv:2611.17688 [pdf, html, other]
-
Title: Exact Failure Probabilities for d-Choice Cuckoo HashingSubjects: Data Structures and Algorithms (cs.DS)
In d-choice cuckoo hashing, each of n items must be placed in one of its d candidate slots in a table of M slots, and every slot holds at most one item. We give a polynomial-time dynamic programming algorithm to compute the exact probability that no valid placement exists. Our method also handles the variant with separate tables.
- [122] arXiv:2611.17689 [pdf, html, other]
-
Title: Orbit-AWARE: From Onboard AI Deployment to Sustainable In-Orbit IntelligenceYunxin Geng, Kaige Zhu, Tongtong Shu, Zhan Gao, Yanxiao Li, Yan Liu, Weiwei Jiang, Zhiqing Wei, Zhiyong FengComments: 10 pages, 4 figures, 2 tables, submitted to IEEE Wireless CommunicationsSubjects: Systems and Control (eess.SY)
The growing role of space computing in time-critical satellite services makes the long-term reliability of onboard artificial intelligence (AI) a mission-level requirement. Model drift can undermine this reliability through critical-observation suppression or noncritical priority flooding, displacing both mission information and evidence for model correction. To address this drift-induced evidence displacement, this article presents Orbit-AWARE, a safeguard-first lifecycle management (LCM) architecture. A comparison of existing approaches motivates cooperative Data, Cognitive, and Trust Planes for continuous monitoring, event-triggered agentic orchestration, and deterministic governance. The architecture retains a model-independent sample stream before degradation is confirmed and coordinates evidence selection and exchange for diagnosis and ground-assisted adaptation. Candidate updates are verified before control over data disposition is restored. In an xBD-based emergency Earth observation (EO) case study, Orbit-AWARE delivers 14.3\% more critical observations and achieves 8.3\% higher sustainable mission utility than the Rule-Based Drift Loop under a common 10\% downlink budget. Finally, we discuss future research directions for extending LCM through task-aware reliability assessment, memory- and tool-enabled onboard agents, and evidence-centric space networking. This work makes evidence availability and recovery authorization explicit design objectives for dependable onboard AI services.
- [123] arXiv:2611.17690 [pdf, html, other]
-
Title: Simulating Space Bounded Read-$k$ PTMs with Space Bounded Read-Once PTMsSubjects: Computational Complexity (cs.CC)
We show that read-$k$ Probabilistic Turing Machines with bounded space $S$ and running time $T$ can be simulated by read-once Probabilistic Turing Machines with error $\varepsilon$, running in space $ O(k^2 (S + \log (n+ T)) + \log (1/\epsilon))$ and time $\frac{1}{\varepsilon} \cdot 2^{O\left(k^2\left(S+\log(n+T) \right)\right)}$. As a direct consequence we show that read-$k$ $\mathsf{BPL}$ is equal to $\mathsf{BPL}$, for any constant $k$. Recently OpenAI announced that $\mathsf{BPL} = \mathsf{L}$. If this is true then our result immediately implies that read-$k$ $\mathsf{BPL}$ is equal to $\mathsf{L}$.
- [124] arXiv:2611.17691 [pdf, html, other]
-
Title: Frequency-Security-Aware Real-Time Economic Dispatch with Data Center UPS Flexibility for Inertia and Primary Frequency Response ServicesSubjects: Systems and Control (eess.SY)
The increasing penetration of renewable energy and the rapid growth of AI data centers are introducing significant challenges to power system frequency security. There is an urgent need to identify grid-compatible resources capable of providing fast frequency support. This paper proposes a frequency-security-aware real-time economic dispatch (RTED) model that enables data center (DC) uninterruptible power supply (UPS) systems to provide steady-state power balancing, inertia response, and primary frequency response (PFR) services. Specifically, tractable analytical frequency-security constraints are derived and incorporated into the RTED framework through second-order cone programming (SOCP) formulation. The proposed model jointly optimizes energy, reserve, and inertia to minimize system operation cost while preventing under-frequency load shedding (UFLS). Numerical validation on a modified IEEE 39-bus system demonstrates that the proposed RTED model with UPS flexibility can economically support system operation and significantly improve frequency security under active-power deficit disturbances.
- [125] arXiv:2611.17693 [pdf, html, other]
-
Title: SPLIT-MOR: Fast and Accurate Model Order Reduction for Nonlinear Transient Thermal Simulation of 2.5D/3D ICsComments: Accepted by ASPDAC 2027; 7 pages, 6 figuresSubjects: Computational Engineering, Finance, and Science (cs.CE)
Fast and accurate transient thermal simulation is essential to thermal reliability analysis and dynamic thermal management of advanced packages, and model order reduction (MOR) techniques are widely applied to overcome the efficiency bottleneck of rigorous numerical methods. However, nonlinear terms including temperature-dependent leakage power, thermal conductivity, and heat capacity prevent the application of conventional linear MOR, while hyper-reduction solutions require costly nonlinear training data. In this work, we propose SPLIT-MOR (Subdomain-wise Piecewise Linearization with Interval Tensors), a novel nonlinear MOR method that requires no additional training set. SPLIT-MOR computes reduced component tensors by subdomain clustering and temperature-interval partitioning in the offline stage, and updates the reduced coefficient matrices and load vectors efficiently in the online stage. Experimental results show that the online solving time of SPLIT-MOR achieves up to an 830x speedup over COMSOL and a 7.5x speedup over the state-of-the-art hyper-reduction method MDEIM, with a 27x total reduction in the offline time. The relative average error remains below 0.11%, and the maximum error is below 0.86 degrees Celsius, demonstrating the high accuracy of our method.
- [126] arXiv:2611.17695 [pdf, html, other]
-
Title: Open-World Scene DeocclusionComments: 10 pages, 7 figures, Accepted at WACV 2027Subjects: Computer Vision and Pattern Recognition (cs.CV)
Recovering the complete appearance of occluded objects from a single image is a fundamental problem for scene understanding and image editing. Real-world scenes are challenging due to clutter, complex spatial occlusions, and diverse object categories. Existing methods often struggle when occlusion involves background surfaces or objects that are not explicitly detected, and fail to resolve the cyclic occlusion dependencies that frequently occur in natural scenes. We introduce a training-free, open-world scene deocclusion framework that recovers the full amodal appearance of detected or user-selected objects from a single RGB image, decomposing the scene into depth-ordered RGBA layers that are geometrically plausible, semantically faithful, and visually consistent with the visible scene. Extensive experiments show that our approach improves scene-level deocclusion, producing coherent and realistic reconstructions in real-world settings. Code will be available on the project page: this https URL
- [127] arXiv:2611.17696 [pdf, html, other]
-
Title: Co-Sight Harness: Externalized Control Assets and Deterministic Enforcement for Long-Horizon Business AgentsLuyao Wang, Xijun Zhu, Zhaohui Hao, Xiwen Zhang, Xingyu Wen, Weishan Li, Yuyang Liao, Yanqin Gao, Yongsheng Du, Lingjun Huang, Shaowei Zhao, Gang Li, Qifan Yu, Baoli WangSubjects: Artificial Intelligence (cs.AI)
Long-horizon business workflows fail in two ways: errors accumulate as a chain of activities lengthens, and the knowledge an activity needs is written in a different vocabulary from the user's request. We study whether such workflows can be made reliable without changing the language model. Our method keeps the backbone fixed and moves the workflow definition, state and domain knowledge into versioned control assets outside the model. A state machine and machine-checkable acceptance gates enforce these assets during execution; an independent auditor grades the resulting trace. The assets are updated only after a regression check and human approval .We evaluate the method on three deployed workflows exceeding 200 steps. On Domain A, a 120-case end-to-end evaluation set shows unattended completion rising from 50% to 100%, retrieval recall from 56% to 91.67%, and end-to-end latency dropping from 440 to 159 minutes, with scenario-analysis recall rising from 45% to 93%. Two other domains were taken over by teams that did not write the assets, one in four days and one in seven days. A controlled ablation separates the two failure sources. The code is available at this https URL.
- [128] arXiv:2611.17700 [pdf, html, other]
-
Title: A Unified Framework for First- and Second-Order Energy-Stable Parametric Finite Element Methods for Planar Willmore FlowSubjects: Numerical Analysis (math.NA)
The construction of structure-preserving schemes for planar Willmore flow that are second-order accurate in time and avoid large coupled nonlinear solves remains challenging. In this work, we develop a unified framework for first- and second-order fully discrete energy-stable parametric finite element schemes for planar Willmore flow with a length-penalization term. The framework is based on a new equivalent reformulation involving a global auxiliary variable. The main novelty lies in the second-order construction, which combines midpoint curvature quantities weighted by elementwise arc-length Jacobian ratios with a compatible discrete tangent to obtain exact discrete chain rules for the bending energy and curve length. A further contribution is to show that both schemes satisfy discrete modified energy dissipation laws that yield the same type of a priori bounds on the discrete solution as the original energy law. They also retain Barrett--Garcke--N{ü}rnberg tangential motion to promote favorable mesh distribution. Each stage reduces to two linear solves with exactly the same coefficient matrix and one scalar nonlinear solve. We rigorously prove unique solvability of the linear subproblems under suitable assumptions and derive computable sufficient conditions for local existence and uniqueness of the scalar solution near unity. Numerical experiments demonstrate the expected temporal accuracy, energy dissipation, good mesh quality and computational efficiency.
- [129] arXiv:2611.17703 [pdf, html, other]
-
Title: iGRAIL: Blockchain-Based Generative Federated Learning for Trustworthy Healthcare IntelligenceComments: 55 pages, 45 figures, 9 tables. PreprintSubjects: Cryptography and Security (cs.CR)
Healthcare organizations collect huge amounts of patient data from clinical records, medical tests, monitoring systems, and healthcare services. Such data can underpin intelligent healthcare analytics but remains distributed across institutions and cannot be shared directly due to privacy, security, and trust concerns. Federated learning (FL) enables collaborative learning without sharing raw patient records, but the exchanged model information remains vulnerable to privacy leakage, tampering, replay, and unauthorized access. In this paper, we propose iGRAIL, a blockchain-enabled generative FL framework for trustworthy healthcare intelligence. iGRAIL combines local generative learning for data augmentation with differential privacy to protect sensitive information during local training. Model communication is secured using cryptographic mechanisms, and blockchain-based verification and smart-contract access control ensure that only authorized, authentic, fresh, and integrity-preserving updates are accepted for federated aggregation. The framework is evaluated on four healthcare datasets and compared with the conventional FL and blockchain-based FL (BC-FL) under predictive, privacy, non-IID, scalability, malicious-update, security-verification, and runtime settings. For the two large diabetes datasets, iGRAIL improves accuracy by about 5.4-5.6%, the F1-score by 14.6-16.9%, and MCC by 16.0-18.1% over the best FL or BC-FL baseline. Increasing the strength of differential privacy reduces the privacy-loss parameter $\epsilon$ by about 72.5% for the cancer dataset and 87.5% for the diabetes dataset, with only a limited loss in prediction performance. These results indicate that iGRAIL supports secure and scalable federated computing while preserving privacy, model integrity, and controlled participation, at the cost of additional computational overhead.
- [130] arXiv:2611.17704 [pdf, html, other]
-
Title: An AI-enabled high-resolution chlorophyll-a forecasting model that outperforms the operational global biogeochemical systemSubjects: Artificial Intelligence (cs.AI); Atmospheric and Oceanic Physics (physics.ao-ph)
Surface chlorophyll-a concentration is a key indicator of marine phytoplankton variability, and short-term forecasts of its distribution provide advance information for marine ecosystem monitoring and management. However, interactions between physical transport and biological processes complicate chlorophyll-a forecasting, and simulating these coupled processes with conventional numerical models imposes substantial computational demands. Here we introduce ChlCast, an AI model that generates daily forecasts of surface chlorophyll-a concentration over a 10-day horizon at approximately 4-km resolution between 60$^\circ$S and 60$^\circ$N. ChlCast combines factorized spatiotemporal convolutions with a multiscale architecture to learn from historical chlorophyll-a fields and climatological information. Compared with the operational 0.25$^\circ$ Copernicus global ocean biogeochemical forecasting system, ChlCast reduces mean absolute error and root mean square error by 70.74% and 47.63%, respectively, against satellite-derived Level-4 chlorophyll-a fields, and by 35.03% and 13.07% against independent BGC-Argo in situ observations. Coastal case studies further show that the 4-km forecasts retain coastal gradients and localized concentration maxima that are weakened by coarse-grid averaging. These results demonstrate the potential of observation-driven deep learning for high-resolution forecasting of surface chlorophyll-a variability at the near-global scale.
- [131] arXiv:2611.17707 [pdf, html, other]
-
Title: RaFL: An Empirical Study of Random Filter Losses for Image ReconstructionSubjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Pixel losses train image reconstruction networks toward a point statistic of the target distribution, often averaging away fine textures under uncertainty. Perceptual losses alleviate this issue using pretrained features, but those features are inherently domain-specific. We study a random filter loss (RaFL) that compares prediction and target through filters with random weights, avoiding both pretrained features and task-specific loss engineering. RaFL combines random convolutions, random Fourier features, and random wavelet filters, automatically balancing their contributions through gradient-norm matching, eliminating manual loss-weight tuning and grid search. We systematically investigate three design choices: filter family, weight resampling schedule, and weight initialization. Based on validation PSNR and stability across random seeds, we select the configuration that uses all three filter families, redraws filters every optimization step, and employs Gaussian initialization. We evaluate RaFL on three 2D and two 3D image reconstruction tasks. RaFL outperforms a pixel loss on three tasks and matches it on two, and demonstrates stable performance across random seeds on all five tasks. RaFL adds less than 2% to the backbone's forward FLOPs and no inference cost, requiring $182\times$ fewer loss FLOPs than a VGG19 perceptual loss. These findings suggest that RaFL is a simple, tuning-free, and domain-agnostic loss objective for 2D and 3D image reconstruction.
- [132] arXiv:2611.17708 [pdf, html, other]
-
Title: QCAP-VQA: Question-Conditioned Capped Prior-Mass Supervision for Medical Visual Question AnsweringSubjects: Computer Vision and Pattern Recognition (cs.CV)
Medical visual question answering models can answer correctly without attending to the anatomy that the question names. Automatic region priors allow spatial supervision without expert masks, but they are noisy, and it is unclear how strongly they should shape attention and how their effect should be verified. We present QCAP-VQA, which supervises a subset of cross-attention heads with a question-conditioned prior built from ranked segmentation proposals. Its capped prior-mass objective penalizes insufficient attention inside the prior only below a threshold. We evaluate attention against SLAKE structure boxes that play no role in training, separating attention on the named anatomy from how attention shifts between questions about the same image. QCAP-VQA matches the accuracy of a baseline without spatial supervision (84.26% vs. 84.23%), and every evaluated prior-supervised model places more attention on the annotated structures. Over a 64-fold range of regularization strength, the capped objective keeps prior mass between 0.67 and 0.77, whereas a pooled region-mass objective drives it to 0.97 and loses accuracy. The cost lies in question conditioning: as the constraint widens or strengthens, attention shifts agree less with the expert-defined contrast. Across 13 configurations, prior-referenced and expert-referenced evaluation rank the models in largely opposite order (Spearman correlation of -0.87), and the capped models it favors track the prior's own errors. Because clinical use of medical visual question answering requires answers grounded in the queried anatomy, supervision from automatic priors should be judged against independent anatomical references.
- [133] arXiv:2611.17710 [pdf, html, other]
-
Title: Cross-Linguistic Acoustic Biomarkers of Cognitive Impairment Across Interactive and Single-Speaker Speech TasksKenyu Ikeuchi, Tomoaki Ohtsuki, Michelle Cohn, Alyssa Weakley, Yui Ishihara, Joseph Hoang, Taishiro Kishimoto, Sarah Tomaszewski Farias, Jet M. J. Vonk, Chen-Nee ChuahComments: 7 pages, 1 figure, 5 tables. Accepted for publication in Proc. 48th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2026)Subjects: Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS)
Multilingual dementia detection is essential for scalable and equitable screening. In recent years, automated speech analysis has emerged as a promising approach for supporting dementia detection by enabling non-invasive and scalable assessment of cognitive function. However, existing studies that use automated speech analysis often rely on one type of task, such as picture description, and use black-box models. These studies often neglect the role of task and language. To address these gaps, we performed a cross-linguistic and cross-task acoustic analysis that integrates a conversational dataset with structured speech tasks across four languages (Spanish, Mandarin, English, Japanese). We applied statistical, linguist-/clinician-guided, and machine learning model-based feature selection to derive compact acoustic feature sets while maintaining classification performance. This enabled systematic comparison of cognitive impairment-relevant features across datasets and tasks. Our results revealed that acoustic features did not generalize uniformly across tasks and languages. Temporal features remained robust across all datasets, consistently capturing timing disruption regardless of language or task. In contrast, speech rate showed strong task dependence -- emerging as informative mainly in single-speaker tasks with externally anchored content and pacing, but becoming less reliable when speech content was highly individualized. Pitch-related features further exhibited language-specific patterns, favoring chroma-related features in tonal languages (e.g., Mandarin) and F0-related features in non-tonal languages. These findings caution against indiscriminate pooling of heterogeneous speech data and underscore the need for interpretable, context-aware approaches to multilingual dementia detection using automated speech analysis.
- [134] arXiv:2611.17713 [pdf, html, other]
-
Title: DesignSpecs: A DESIGN.md Dataset of Visual Design Specifications for AI AgentsSubjects: Software Engineering (cs.SE); Human-Computer Interaction (cs.HC)
While AI agents increasingly generate and modify user interfaces, they rarely rely on a visual design specification for a certain project. The DESIGN$.$md format, which originated from Google's AI design tool Stitch and which was open-sourced in April 2026, addresses this gap by describing a visual identity through design tokens (such as named colors, text styles) and Markdown prose. To our knowledge, no large-scale research dataset of such design specification files exists. We present DesignSpecs, a dataset of 36,906 DESIGN$.$md files from 15,794 public, non-fork GitHub repositories, representing 17,240 unique file contents. We collected 613,207 files named DESIGN$.$md in October 2026. Because the same file name also denotes architectural plans and other design documents, we filtered those with YAML front matter, a design-system name, at least one recognized design token, and zero errors under the official linter. Our JSON Lines dataset provides file-level and deduplicated views of the original Markdown, links files to repositories, paths, and commits, and is released with related scripts. We outline research opportunities on visual design knowledge and design-aware AI agents. Dataset is available at: this https URL
- [135] arXiv:2611.17720 [pdf, html, other]
-
Title: Ego2Flow: Benchmarking the Traffic-System Impact of Autonomous Driving on Maps Distilled from Real-World RoadsYu Dong, Weiying Shen, Hangyu Li, Shu Han, Hong Li, Xiangbo Gao, Keshu Wu, Zhengzhong Tu, Soyoung Ahn, Zhiyuan Liu, Xiaopeng LiSubjects: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
End-to-end autonomous driving is predominantly evaluated through ego-vehicle safety and task completion. Its deployment in urban traffic, however, creates sustained interactions among vehicles whose collective effects remain insufficiently evaluated. We introduce Ego2Flow, a closed-loop benchmark for assessing the traffic-system impact of autonomous driving on maps distilled from real-world cities. To construct realistic 3D urban environments for sustained closed-loop traffic simulation, we develop an agent-guided map distillation framework that integrates real-world maps and observations with complementary geographic information. Through an iterative feedback mechanism and a traffic information graph that jointly guide scene generation, the framework preserves lane topology, critical road geometry and decision-relevant visual conditions while simplifying nonessential details and completing unobserved regions. Using our proposed framework, we generate and publicly release six CARLA maps from the Argoverse 2 dataset, covering the corresponding real-world regions and enabling paired comparisons of driving behavior under real and simulated observations. On these maps, Ego2Flow evaluates how autonomous driving policies affect traffic efficiency, safety, stability and energy consumption across diverse road structures. Compared with existing construction methods, our framework achieves the lowest average displacement error between planning outputs from paired real and simulated observations. We further benchmark representative driving policies to characterize their system-level impact on urban traffic. Project Page: this https URL.
- [136] arXiv:2611.17722 [pdf, html, other]
-
Title: Federated Soft Poly-Level Learning Over a Distributed Network with Guaranteed ConvergenceComments: Accepted at IEEE Transactions on NetworkingSubjects: Artificial Intelligence (cs.AI)
Multi-level learning/optimization, e.g., bilevel optimization, has been recognized for its versatile applicability across a broad array of domains, spanning from signal processing to machine learning. However, most existing works focus on bilevel and trilevel optimization. In contrast, multi-level optimization problems involving four or more hierarchical levels, which we term poly-level learning (PLL) problems, remain largely this http URL addition, in practical PLL, data may be distributed across multiple nodes over a distributed network. Nevertheless, extant studies on PLL predominantly focus on the non-federated setting, necessitating the collection of data to central servers and thereby posing potential privacy risks. To this end, we propose a Federated Poly-Level Learning (FPLL) framework over a distributed network to tackle the PLL problems with soft constraints in a federated manner. FPLL hinges on a novel non-linear cut, i.e., $L$-cut, for the level-wise relaxation to the PLL problems. Additionally, the proposed federated algorithm is designed with adjustable parameters that allow for a flexible trade-off between its computational complexity and performance efficiency, enabling users to tune it according to specific application requirements. We also carry out non-asymptotic analysis to derive the iteration complexity for the proposed FPLL to achieve the $\epsilon$-Pareto stationary point. Extensive experiments have been undertaken to elucidate the superior efficacy of the proposed method.
- [137] arXiv:2611.17723 [pdf, html, other]
-
Title: uhd-rfsoc: Bringing Off-the-Shelf RFSoC Boards to Unmodified UHD ApplicationsSubjects: Networking and Internet Architecture (cs.NI); Hardware Architecture (cs.AR); Signal Processing (eess.SP)
Next-generation wireless research relies on software-defined radios (SDRs) with more channels and larger bandwidth. The USRP family is one of the most widely used SDR platforms, and much of the wireless software ecosystem is built on top of the USRP hardware driver (UHD). However, each USRP model comes with a fixed channel count, bandwidth, and streaming capability. Off-the-shelf radio frequency system-on-chip (RFSoC) boards, on the other hand, integrate up to 16 analog-to-digital converters (ADCs) and 16 digital-to-analog converters (DACs) operating at several gigasamples per second (GS/s), but existing UHD applications cannot directly use them. We present uhd-rfsoc, an open-source FPGA design and host software stack that turns these boards, including the 16-channel ZCU216 and the RFSoC 4x2, into standard UHD devices without modifying UHD. Through standard UHD calls, applications can configure the carrier frequency, channel count, the per-channel I/Q sample rate, and the wire format. uhd-rfsoc combines timestamped, flow-controlled sample transport in the FPGA with data plane development kit (DPDK) packet processing and vectorized sample conversion on x86 and Arm hosts. The built bitstreams for a number of RFSoC board configurations with varying RF data converter rates, channel counts, I/Q sample rates, and wire formats all successfully pass UHD's built-in streaming benchmarks. We further demonstrate compatibility with unmodified GNU Radio flowgraphs, a custom GPU-accelerated transceiver digital signal processing pipeline, as well as the OCUDU and Duranta 5G radio access network stacks using both software and commodity user equipment.
- [138] arXiv:2611.17726 [pdf, html, other]
-
Title: Towards High-Performance Multimodal Discovery: Evaluating Retrieval Efficacy and Operational Scalability in Digital LibrariesSubjects: Information Retrieval (cs.IR)
Large-scale digital repositories with million plus digitized volumes serve as vital engines for scientific and scholarly discovery. However, as collections grow, finding specific research or educational material remains a significant challenge. This difficulty is compounded by the multimodal nature of texts, that is, texts which combine dense narrative content with essential illustrative charts, annotated diagrams, and complex layouts. Historically, discovery has been supported by full-text search indexes implemented as inverted indexes (e.g., Lucene or Solr). While these systems scale well for textual strings, they are fundamentally "blind" to the semantic nuance of visual information. We explore the benefit of discovery through texts that have been converted to numerical vectors. By representing page images as numerical vectors rather than OCR-extracted text, and doing so for both text and images on a page, we capture both textual and visual semantics while mitigating reverse engineering and copyright infringement concerns. The primary contribution of this research is an evaluation of the multimodal retrieval landscape, encompassing both semantic precision and architectural scalability. Our findings demonstrate that integrating visual and textual features significantly enhances the discovery of relevant material, particularly in domains where diagrammatic content is essential for conceptual understanding. We assess retrieval efficacy, demonstrating that the system achieves sub-linear scalability. Our findings suggest that an optimized vector database configuration is the critical factor in transforming VLM-based embeddings into a production-ready, library-scale discovery solution.
- [139] arXiv:2611.17728 [pdf, html, other]
-
Title: A Spectral Framework for Universal Proxy Regularization in Model MergingSubjects: Artificial Intelligence (cs.AI)
Model merging constructs multi-task models without expensive joint training but often suffers from parameter interference. Some solutions mitigate this by constraining finetuning using data from the other tasks to be merged, violating the core independence and privacy assumption of model merging. Recent work offers preliminary empirical evidence that a universal, task-agnostic proxy distribution can substitute for such data, but leaves open when and why it suffices. To this end, we propose a \textit{Spectral Framework for Universal Proxy Regularization}, which characterizes the relationship between a proxy distribution and the distribution of other tasks to be merged through spectral coverage ($\alpha$) and leakage ($\beta$). Based on the framework, we derive a unified parameter interference bound: smaller $(\alpha, \beta)$ provably yields lower parameter interference through better spectral alignment. To exploit such a proxy, we augment the training objective with a proxy regularizer. Evaluating this regularizer during fine-tuning requires a concrete mechanism. We study three instantiations, including (J) an exact evaluation via Jacobian-vector products, (F) a frozen curvature approximation using the Fisher information matrix, and (O) an output-drift formulation that constrains the model's functional deviation on the proxy directly. We prove that all three coincide under the NTK linearization, where existing analyses of such regularizers stop, and go further to characterize standard non-linear fine-tuning, bounding how far each instantiation departs from its linearized reference and obtain a three-way hierarchy, zero for (J), proxy-independent for (F), and proxy-quality-scaling for (O). Empirical validation on the commonly used vision-language model CLIP confirms these theoretical predictions.
- [140] arXiv:2611.17730 [pdf, html, other]
-
Title: ERA: Error-Resistant Autonomous Object Navigation with Predicted DepthSubjects: Robotics (cs.RO)
Depth and 3D foundation models can predict geometry from RGB images, allowing zero-shot object navigation without a depth sensor. However, when predicted depth is mapped as if it came from a sensor, view-dependent errors accumulate in the map, remove exploration goals, and end the search before the agent can observe and correct them. We propose ERA, a training-free object navigation system that replaces the depth sensor with depth predicted from RGB images and poses. ERA follows a paradigm we call Move with Error: movement provides new views that correct map errors, and the corrected map enables further movement. It realizes this paradigm with revisable occupancy and with exploration from reachable viewpoints, which continues the search when no frontier remains, and verifies the target across views before stopping. On three benchmarks, ERA achieves state-of-the-art performance among training-free methods without depth sensors. Its map and exploration components also raise VLFM's success from 10.8% to 61.7% under the same predicted depth. We further deploy ERA on a real robot.
- [141] arXiv:2611.17733 [pdf, html, other]
-
Title: HarmonyText: Harmonious Scene Text Editing via Reverse-to-Refined Pair Synthesis and Region-Decoupled Frequency LearningZiqi Yi, Huiguo He, Guitao Xu, Peirong Zhang, Cheng Da, Xinyao Yu, Zhanpeng Wang, Changqian Yu, Huan Yang, Kun Gai, Lianwen JinComments: 11 pages, 7 figures. Accepted to SIGGRAPH Asia 2026 Conference PapersSubjects: Computer Vision and Pattern Recognition (cs.CV)
Natural language-guided, mask-free scene text editing aims to modify text in real-world images using only textual instructions, yet remains challenged by unreliable pseudo-GT supervision, over-smoothed text generation, and the absence of GT-free evaluation. Existing methods often rely on imperfect pseudo-targets with background discrepancies, while standard MSE-based velocity regression tends to suppress high-frequency details, leading to blurred text boundaries and weakened local textures. To address these issues, we propose HarmonyText, a mask-free scene text editing framework with region- and frequency-aware supervision. HarmonyText introduces Region-Decoupling Optimization (RDO), which aligns edited text regions with pseudo-targets while anchoring preserved regions to the source image, and Cascaded High-Frequency Texture Guidance (CHFTG), which applies cascaded wavelet-domain constraints to preserve sharp boundaries and fine-grained textures of textual content. We further propose Reverse-to-Refined Pair Synthesis (RePair) to construct STE25K, a high-quality 25K paired dataset, and introduce a GT-free evaluation protocol with Color Distribution Divergence and a Harmony Assessment Model. Extensive experiments show that HarmonyText outperforms existing methods in text editing accuracy, background fidelity, color consistency, and perceptual harmony.
- [142] arXiv:2611.17735 [pdf, html, other]
-
Title: Logically Consistent Accurate ForecastsComments: NeurIPS 2026 Workshop on Foundation Models for Temporal Systems (FMTS)Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Agentic forecasting systems based on large language models (LLMs) are now competitive with expert human forecasters, but these systems still suffer from a critical limitation: the inability to produce forecasts that are logically consistent. For example, forecasts for related events often fail to satisfy the axioms of probability, which in turn implies that related forecasts cannot simultaneously be correct and undermines their utility for decision makers. Enforcing consistency is not trivial -- naive approaches (e.g., projection onto the set of consistent forecasts) can degrade accuracy. We argue that the key to overcoming the apparent tradeoff between consistency and accuracy lies in the reasoning which supports each forecast. We finetune a supervisory language model that leverages this reasoning to jointly reconcile inconsistencies while improving the accuracy of related forecasts. On ForecastBench, our method improves the Brier score of a retrieval-augmented GPT-5.1 forecaster by 9.3% while reducing measures of structural inconsistency by 82-90%. We interrogate our method via a comprehensive series of baselines and ablations, and find that no alternative method matches our gains in consistency and accuracy simultaneously.
- [143] arXiv:2611.17737 [pdf, html, other]
-
Title: Super-Analysis for Aggressive yet Scalable Compiler OptimizationSubjects: Programming Languages (cs.PL); Symbolic Computation (cs.SC)
Aggressive compiler optimization today occupies two distant extremes. Conventional optimizers run in milliseconds but reason through ad-hoc heuristics that miss opportunities; superoptimizers synthesize faster code but take hours, confining them to small kernels and offline use. We introduce solver-based super-analysis, a middle ground that uses expensive SMT-based analysis while transformations remain conventional, avoiding program synthesis entirely. The key argument for super-analysis is that conventional compilers have rich transformation rules, but proving them to be safe requires complex reasoning: proving a safety check dead often needs facts from three independent LLVM analyses simultaneously, a conjunctive entailment problem suited to an SMT solver, not to any single dataflow pass.
We instantiate super-analysis in ODeSSy, an LLVM pass that proves sanitizer traps and native-language safety checks unreachable and removes them. One mechanism serves four frontend languages and yields speedups in each, including a 3.78x speedup for a GEMM implemented in Julia. A further solver-based optimization, guard synthesis, handles checks that are not dead for every input: the solver synthesizes a cheap invariant safety guard under which a fast check-free loop is cloned, speeding up a Swift Base64 encoder by 1.44x and a Julia LZ77 compressor by 4.01x. Compared with Souper, the synthesizing superoptimizer for LLVM, ODeSSy removes more safety checks in significantly less time, making it viable as an online compiler pass. - [144] arXiv:2611.17739 [pdf, html, other]
-
Title: State-of-Charge Targets for Storage Participation in Two-Settlement Markets: Reliability, Price, and Compensation ImpactsSubjects: Systems and Control (eess.SY)
As energy storage assumes a larger role in electricity systems, market operation must coordinate its use across time to support reliability during periods of system stress. Minimum state-of-charge (SOC) targets provide a means of preserving energy for these periods, but their design shapes reliability, operating efficiency, and storage compensation. This paper evaluates alternative SOC target designs in a California-based day-ahead and real-time electricity market model. The comparison examines targets based on day-ahead schedules or rolling centralized dispatch, enforced at the individual-resource or system level, across seasons, stress conditions, and participant information settings. The results show that retaining more energy does not necessarily reduce load shedding: targets that restrict discharge can limit storage support during shortages. Under critical summer conditions, individual targets derived from rolling centralized dispatch reduce weekly load shedding by approximately 83--84\% relative to operation without targets. Lower RT clearing costs and average prices can coexist with higher storage compensation, reflecting the role of day-ahead positions and real-time deviations in determining settlement outcomes. These findings highlight the importance of coordinating energy retention with timely discharge and network deliverability when designing SOC requirements for storage participation in electricity markets.
- [145] arXiv:2611.17742 [pdf, html, other]
-
Title: Analysing Busy Beaver Machines (Or How to Find Ossified Ocelots)Subjects: Formal Languages and Automata Theory (cs.FL); Logic in Computer Science (cs.LO)
The busy beaver problem involves working with machines which can produce very large numbers, both for the output of certain Turing machines and for the number of steps taken to execute them. Given that these numbers can involve thousands of digits, it is crucial to have a means of proving properties of the relevant machines which allow manipulation of such large numbers. In this paper we introduce the notion of \textit{shape computations} as a means of isolating key properties that allow efficient execution of the relevant Turing machines based on verifiable properties of state transitions. In particular, this notion allows us to provide an analytical result which demonstrates the quadratic relationship between the size of the output and the number of steps required to generate it. We have also developed an initial implementation of our approach, and we provide empirical evaluation of it on a suite of 240 example machines.
- [146] arXiv:2611.17746 [pdf, html, other]
-
Title: Complex Schrödinger BridgesComments: NeurIPS 2026Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Diffusion bridges have emerged as versatile frameworks for facilitating transformations between arbitrary data distributions. However, these approaches are hindered by an immense optimization space and its underlying high-order structures. This paper explores a variant of diffusion bridges for solving real-world problems in complex vector spaces governed by scaling and rotation laws, such as Hamiltonian structures. To address this, we introduce conformal optimal transport (COT), a complex-valued transportation framework designed for physical systems with geometric and dissipative dynamics under a geometric lifting. Building on this, we propose a new class of dynamical diffusion bridges, called complex Schrödinger bridges (CSBs), whose derived algorithms can handle diffusion-based tasks using a conformal geometric prior. For scalability, we propose image-to-image CSB (I$^2$CSB), a tractable diffusion-bridge technique with an analytic spiral posterior. Finally, we implement a physics-informed CSB algorithm within simulations of quantum electrodynamics, mimicking the logical behavior of quantum logic gates. Our extensive results validate the underlying theory while highlighting advantages of complex-valued diffusion.
- [147] arXiv:2611.17747 [pdf, html, other]
-
Title: Evaluating the Robustness of PRNU Source-Camera Attribution to the Fingerprint-Copy Attack: An Open Toolkit with Triangle-Test DetectionComments: 8 pages, 3 figures. Toolkit: this https URLSubjects: Cryptography and Security (cs.CR)
Photo-Response Non-Uniformity (PRNU) is the de facto standard trace for sensor-level source-camera attribution, and a high Peak-to-Correlation-Energy (PCE) between a questioned image and a camera reference fingerprint is routinely treated as strong evidence of common origin. The fingerprint-copy attack of Goljan, Fridrich and Chen (2011) shows that this signal can be forged: an adversary who estimates a target camera's fingerprint from that camera's publicly available images can inject a scaled copy of it into an unrelated photograph, producing a forgery that clears a PCE decision threshold despite never having been taken by the target device. We present prnucf, an open, auditable toolkit that pairs a reference implementation of the (additive) fingerprint-copy attack with the triangle-test detector and a cohort-calibrated decision procedure, so that practitioners can quantify how fragile a PRNU attribution is against this attack on their own pipelines. On a controlled multi-camera Monte-Carlo benchmark the attack reaches an arbitrary target PCE at a mean PSNR of 59.2 dB (visually lossless), yet the triangle test separates forgeries from innocent images with an area under the ROC curve of 0.999, detecting every forgery at a cohort-calibrated 5% false-positive operating point. We discuss why a fixed significance cut-off is not safe in casework, the limitations of synthetic validation, and the open problem of extending explainable detection to the modern, GAN-based generation of fingerprint-transfer attacks. The toolkit is released under the MIT license with an explicit dual-use statement.
- [148] arXiv:2611.17751 [pdf, html, other]
-
Title: When Does Cross-View Consistency Matter in Multi-View Appearance Augmentation? A Controlled Study for Visuomotor PoliciesComments: 30 pages, 10 figures. Code: this https URLSubjects: Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
Multi-view visual augmentation often seeks consistency across cameras, but it is unclear which forms of consistency matter for downstream control. We use simulation-state replay to manipulate cross-view coupling while preserving the per-camera perturbation distributions, physical trajectories, and action labels. Across seven LIBERO manipulation tasks, marginal-matched changes in global appearance, ghost injection, and target color have small average costs for a late-fusion Diffusion Policy: point estimates are near zero or negative, and the largest upper endpoint of the 90% posterior credible intervals is 2.55 percentage points. In contrast, geometrically anti-coupled observations reduce success by 3.30 percentage points relative to coupled observations (90% CrI [0.64, 6.28]), with displacement magnitude, direction marginals, and the number of affected views matched. Camera interventions reveal substantial dependence on wrist-view geometry, while feature analysis shows appearance sensitivity patterns associated with the factors randomized during training. ACT shows a higher target-color cost estimate than DP, with uncertainty that includes zero and sensitivity to task composition. Within the tested tasks and policy configurations, these findings motivate selective consistency constraints: appearance coupling can sometimes be relaxed, whereas task-relevant geometry and cross-view spatial correspondence warrant protection.
- [149] arXiv:2611.17752 [pdf, html, other]
-
Title: Stage-wise Lineage-aware Dropout for Sparse-View 3D Gaussian SplattingComments: 11 pages, 9 figures, 7 tables, including supplementary materialSubjects: Computer Vision and Pattern Recognition (cs.CV)
Sparse-view 3D Gaussian Splatting (3DGS) is prone to overfitting because only a limited number of images are available to constrain scene geometry. Dropout is one regularization strategy for alleviating this problem. We introduce the concept of a lineage family to describe a set of Gaussian primitives derived from the same founder through cloning or splitting operations. The founder Gaussians refer to all Gaussian primitives generated during initialization. Descendants within the same family often jointly contribute to the refinement of adjacent local regions. However, dropping individual Gaussians can be less effective due to the compensation effect from neighboring Gaussians for the dropped ones. To address this issue, we propose Stage-wise Lineage-aware Dropout(SLD), which applies lineage-family dropout during the structural regularization stage and switches to individual dropout during the later refinement stage. We further propose Lineage-Balanced Densification(LBD), which operates during densification by constructing an expanded candidate pool and reallocating the densification budget according to the mean gradient scores and existing descendant sizes of different families. Experiments on the LLFF, MipNeRF360, and Blender datasets show that the proposed framework achieves competitive performance in sparse-view reconstruction
- [150] arXiv:2611.17753 [pdf, html, other]
-
Title: Superfluid: Local AI Server for Distributed LLM InferencePrabod Rathnayaka, Isura Manchanayake, Srikandabala Kogul, Steve Siu, Fabian Waschkowski, Lukas WesemannSubjects: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Running large language models (LLMs) on local devices keeps data private, removes per-token cost, cuts network latency, and keeps working without a connection. Two shifts now strain how such models are served. On one device, several agents share a locally served model with an interactive user. Across devices, a developer or small team often owns several capable machines from different vendors, which sit idle most of the day because every local server serves one runtime on one machine. We present Superfluid, an open-source LLM server built for local inference at both scales. On one device, Superfluid drives this http URL, MLX, or BaseRT in isolated worker processes through one scheduler that moves interactive requests ahead of agent work, computes a prompt shared by many requests once, and writes every token to a durable log before it reaches the client. Across devices, one Superfluid server turns a fleet of machines into a single API endpoint. Machines with different hardware serve copies of the same model, receive work according to their measured speed, and can join or fail under load without losing a request. For Qwen3.8-27B on an Apple M5 Pro, a chat sent while eight agents are working receives its first token in 1.5 s, against 291-612 s on llama-server, Ollama, and this http URL. On six machines from three chip vendors serving 64 simulated users for one hour, Superfluid answers all 5,041 requests while machines join one at a time and one is killed and rejoins; throughput grows from 112 to 534 tok/s and the median first-token time falls from 99 to 12 s. On twelve Mac Studio M3 Ultras it sustains 10,656 tok/s across 38,563 requests from 384 agents without a failure. Superfluid is available under the Apache-2.0 licence at this https URL.
- [151] arXiv:2611.17754 [pdf, html, other]
-
Title: See Every Block Once: Variance-Reduced Post-Training for Masked Diffusion Language ModelsComments: AACL-IJCNLP 2026 MainSubjects: Computation and Language (cs.CL)
Policy gradient methods are widely used to align large language model outputs with human preferences, but the log-likelihoods they require cannot be directly computed in Masked Diffusion Language Models (MDLMs). MDLM post-training therefore relies on a surrogate objective based on the Evidence Lower Bound (ELBO), and recent work estimates a blockwise variant of this objective via Monte Carlo methods to support blockwise generation. However, existing estimation schemes randomly sample the block index, so they may evaluate the same block more than once and leave some blocks unevaluated. This introduces variance arising from block selection into the estimates. In this work, we propose a full-block estimator that evaluates every block exactly once and averages the results. We theoretically show that the proposed estimator remains unbiased while eliminating the between-group variance arising from block selection, thereby achieving lower variance than the baseline for the same sample size. In experiments on five benchmarks spanning mathematics, planning, and coding, our method achieves the highest accuracy on four benchmarks and outperforms the random block sampling baseline by an average of 12.79 points.
- [152] arXiv:2611.17755 [pdf, html, other]
-
Title: VACO: Vertex-Aligned Decoding for Dense Hand Contact Estimation from a Single ImageSubjects: Computer Vision and Pattern Recognition (cs.CV)
Dense hand contact estimation from a single RGB image is a foundational primitive for dexterous manipulation and 3D interaction analysis. Direct predictors typically decode the entire per-vertex contact map from one shared query, leaving individual vertices ungrounded in the image and smearing contact across adjacent, non-touching surfaces. We present VACO, a vertex-aligned decoder in which each MANO vertex samples local image features at its projected location. A hand-reconstruction backbone, jointly trained for reconstruction and contact, predicts these locations at inference without an auxiliary reconstructor. Training uses noise-perturbed ground-truth projections and end-to-end feature adaptation to improve robustness and contact-sensitive features. A single VACO checkpoint sets a new state of the art on five benchmarks spanning hand-object, hand-human, hand-scene, and bimanual interactions. It improves F1 by +20.7 pp over the strongest prior direct predictor on Hi4D and generalizes to the held-out HOT3D and HOCap, with gains of +5.5 and +2.7 pp. The resulting contact predictions also improve downstream HOI reconstruction.
- [153] arXiv:2611.17756 [pdf, html, other]
-
Title: Compiling Matrix Formulas to Quantum Block EncodingsComments: 20 pages, 12 figuresSubjects: Programming Languages (cs.PL); Quantum Physics (quant-ph)
Realizing the promise of quantum applications such as chemistry or fluid dynamics requires building quantum circuits that efficiently encode and manipulate exponentially large matrices. This process forces researchers to grapple with the reality that a matrix expressible as one equation on paper can map to thousands of bit-level quantum logic gates, historically written by hand. Even worse, it often ends with a discovery that the needed resources grow so quickly with matrix size that the quantum algorithm loses its asymptotic speedup.
In this work, we present MARBLE, a language that automatically synthesizes an efficient circuit from the publication-style formula for the entries of a matrix. MARBLE algebraically rewrites programs into symbolic combinations of simpler matrices that have known circuit realizations. It analyzes costs of possible representations, notably the subnormalization term that dominates time complexity. It then emits an OpenQASM circuit and certifies whether its cost is poly-logarithmic in problem size, as required by quantum algorithms.
When evaluated on quantum applications ranging from fluid dynamics to physical Hamiltonians, MARBLE matches or surpasses 37 of 39 published baselines. On the low end, it automatically matches many theoretically optimal circuits, and on the high end, it improves on a published construction in fluid dynamics by 10,487x in subnormalization. MARBLE produces exponentially smaller circuits than the leading tool FABLE, and it successfully compiles matrices whose dimensions exceed 2^100 within five seconds on a laptop. - [154] arXiv:2611.17757 [pdf, other]
-
Title: Graph Neural Network Prediction of Frequency Nadir and RoCoF Under Topology VariationsSubjects: Systems and Control (eess.SY)
Accurate prediction of frequency dynamics under varying network topologies is essential for rapid frequency security assessment, whereas repeated electromagnetic transient (EMT) simulations are computationally intensive. This paper develops a topology-aware graph neural network with a radial basis function Kolmogorov-Arnold network (RBF-KAN) regression head, termed RK-GNN, to predict frequency nadir and maximum absolute rate of change of frequency (RoCoF). Scenarios combine load scaling, regional load redistribution, and generator redispatch with intact, N-1, and N-2 transmission topologies. After checks for network connectivity, generator reference limits, and numerical consistency, the retained scenarios are simulated in PSCAD to construct the reference dataset. The RK-GNN is evaluated against a conventional deep neural network (DNN) and a baseline GNN. On the IEEE 39-bus system, the RK-GNN reduces the mean absolute errors (MAEs) of nadir and RoCoF by 48% and 59% relative to the DNN and by 42% and 61% relative to the baseline GNN. Incorporating the bridge indicator and disturbance hop distance yields further MAE reductions of 27% and 54%. These results show that the proposed topology-aware architecture provides accurate prediction of both frequency metrics under varying topologies and substantially reduces the need for repeated EMT simulations.
- [155] arXiv:2611.17758 [pdf, html, other]
-
Title: Efficient Genotype Classification from 3D MRI in ADPKD via Density-Aware Point Sampling and E(3)-invarianceSubjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Early detection of genetic kidney diseases, especially autosomal dominant polycystic kidney disease (ADPKD), is crucial because certain genotypes such as truncating PKD1 mutations cause rapid kidney growth and early progression to end-stage kidney disease. MRI-based genotype inference is a potential low-cost complement to genetic testing. However, although many deep learning methods have been developed for classifying various MR images, applying them to ADPKD remains challenging: (1) abdominal MR images are generally too large to be handled by deep learning models designed for smaller images like brain MRI, and (2) the amount of labeled data for ADPKD genetic mutations is insufficient to train deep learning models. In order to address these challenges, we propose E3PN, an efficient 3D MRI-based ADPKD genotype classification network that relies on point sampling and E(3)-invariance. We utilize a point representation of 3D MRI, which consumes less GPU memory than a voxel representation. For further memory efficiency, 3D points are sampled from a mixture distribution over the two kidneys in each image. Since E(3)-invariance has been shown to improve data efficiency, we also incorporate it into our model through coordinate canonicalization. On our ADPKD dataset, E3PN trains 163 times faster and uses 103x less GPU memory than a 3D ResNet, while also improving F1 from 0.165 to 0.226 over both baselines. These results establish a first benchmark for five-way ADPKD genotype classification from 3D MRI and show that point sampling and coordinate canonicalization yield two-orders-of-magnitude efficiency gains with a modest but consistent improvement in F1.
- [156] arXiv:2611.17759 [pdf, html, other]
-
Title: Convergence analysis of physics-informed neural networks for subdiffusionSubjects: Numerical Analysis (math.NA)
Physics-informed neural networks have emerged as mesh-free solvers for partial differential equations, yet their rigorous convergence theory for time-fractional problems remains limited. We present a convergence analysis of PINNs for linear subdiffusion equations and for the associated backward problem of recovering the initial data from terminal observations. To address the nonlocality and weakly singular kernels of fractional derivatives, we introduce a discretization-based PINN framework built on some popular time-stepping schemes satisfying discrete maximal regularity, including backward Euler, BDF2, L1, and fractional CN schemes. This approach avoids the exact evaluation of fractional derivatives at training points while preserving the stability structure needed for convergence analysis. For the forward problem, we prove that, under the natural fractional maximal-regularity condition $u\in W^{\alpha,p}(0,T;L^2(\Omega))\cap L^p(0,T;H^2(\Omega)\cap H_0^1(\Omega))$, the piecewise linear temporal reconstructions of PINN-loss minimizers converge to the exact solution. This result avoids the higher-order regularity assumptions on the exact solution that are often imposed in existing PINN convergence analyses. We also verify the required discrete maximal regularity and consistency properties for the aforementioned popular time discretizations. We then extend the analysis to the backward subdiffusion problem and establish convergence of the regularized recovery of the initial data from terminal observations, despite its ill-posedness. Numerical experiments for both forward and backward problems support the theoretical results, demonstrate the necessity of regularization, and show that $L^2$-type losses alone are generally insufficient. These results establish a rigorous theoretical basis for discretization-based PINNs for time-fractional diffusion problems.
- [157] arXiv:2611.17761 [pdf, html, other]
-
Title: Inferring Sensitive Events in Multi-Mode Cyber-Physical Systems via Scheduling Side ChannelsSubjects: Cryptography and Security (cs.CR); Systems and Control (eess.SY)
Many cyber-physical systems (CPS) are multi-mode: they switch between operating modes with different tasksets and timing parameters in response to physical events or user activity. We show that these mode changes create a timing side channel: by altering the interference experienced by co-running tasks, they let an unprivileged co-resident observer infer sensitive events, such as a homeowner's arrival or a vehicle's route, from its own execution timing and, on multicore platforms, its core assignment. Existing schedule-based attacks and defenses miss this channel because they assume single-mode systems or target a specific, known victim task. We present STAMP, a lightweight attack that needs no knowledge of task parameters or scheduling policy, and works whether or not the system exposes its modes. On a ZoneMinder home-security deployment, STAMP identifies the modes with 94.7% accuracy on unseen hardware, using under 0.5% of one core. On ROS 2 AMCL with the Oxford RobotCar dataset, it identifies which of seven route segments a vehicle traversed with 86.0% accuracy on unseen platforms. A schedule-randomization defense lowers detection recall to 0.53, at which setting only 53.9% as many jobs finish on time. At higher noise ($\epsilon=10$), on-time job completion reduces to 14.4%, while introducing new leakage in 14.1% of the tasksets.
- [158] arXiv:2611.17762 [pdf, html, other]
-
Title: Creative Product Design under Brand Image ConstraintsSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Graphics (cs.GR)
Industrial design requires creative visual exploration while satisfying constraints that are often product-specific and implicit. Existing text-to-image (T2I) approaches can generate creative designs through concept blending or multi-agent self-enhancement, but they provide limited mechanisms for ensuring that generated designs remain consistent with such constraints. We focus on brand image as a representative design constraint and propose a framework for brand-aware redesign generation. Given a product name, the framework constructs a brand image survey report that summarizes product and market information, derives a creative direction from the report, and integrates this direction into a self-enhancement loop in which generated images are critiqued and prompts are revised. We evaluate the framework on 50 redesign tasks across five product categories, comparing it with Naive T2I, IP-Adapter/IT-Blender, and unconstrained self-enhancement under the same T2I backbone. By a VLM-as-Judge protocol to measure creativity and brand identifiability, the proposed framework achieves a more favorable trade-off between the two objectives, with particularly clear gains in highly constrained categories such as automotive and motorcycle, where the plausible design space is relatively limited compared with that of other product categories. These results suggest that dynamically eliciting and textualizing brand image can guide creative product design generation under brand image constraints.
- [159] arXiv:2611.17763 [pdf, html, other]
-
Title: Knowledge After Collapse: Context and Wording Shape Factual Accessibility in Language ModelsSubjects: Artificial Intelligence (cs.AI)
When a language model accepts a false correction in a factual discussion, its next answer can depend on how the conversation continues. We introduce Knowledge Answer Divergence (KAD), an event-conditioned evaluation protocol that distinguishes initial false-answer adoption from subsequent persistence, recovery, and abstention. KAD freezes the conversation at the first explicit adoption of a designated false target and branches independent probes from that exact history. This enables matched comparisons of post-adoption behavior and history-management interventions that a single susceptibility score cannot capture. We evaluate eight local checkpoint conditions and six hosted endpoints, alongside an audited 300-question rerun. In corrected Llama 3.2 Instruct 3B controls, a neutral-history branch yields 42/46 correct responses, compared with 16/46 under full history. A matched cue-placement control changes false adoption by 27.3 percentage points in Qwen 3.5 4B. Twelve independently authored pressure templates demonstrate substantial sensitivity to wording. Together, these findings demonstrate the utility of KAD for evaluating factual reliability beyond initial susceptibility, enabling systematic comparisons of false-answer persistence, recovery conditions and conversational interventions.
- [160] arXiv:2611.17765 [pdf, html, other]
-
Title: Toward the Minimal Set of Information Inequalities in Regenerating Codes: An Algebraic ApproachSubjects: Information Theory (cs.IT)
Information-theoretic outer bounds for exact-repair regenerating codes can be formulated as linear programs, but the associated Shannon-type systems contain many implicit equalities and redundant inequalities. We directly specify a reduced system for the $(n,n-1,n-1)$ exact-repair model using the closure structure induced by its repair dependencies. Nonempty closed sets index the reduced joint-entropy coordinates, and inseparable edges determine the defining inequalities. We prove equivalence with the initial system by characterizing its implicit equalities and establishing sufficient redundancy criteria. We also give an algorithm that constructs the directly defined system for any $n$, count its entropy coordinates, and derive an explicitly computable upper bound on its number of inequalities. For every $n\ge2$, we prove that the reduced system has no implicit equalities and that each distinct defining inequality is nonredundant and defines a facet in the nonempty closed-set coordinate space. Exact enumerations for $n\in\{2,3,4,5\}$ verify the reported system sizes.
- [161] arXiv:2611.17766 [pdf, html, other]
-
Title: GDP.xlsx: Benchmarking Frontier Models on Professional WorkbooksComments: 14 pages, 4 figures, 6 tables. Accepted to the Agent Evaluation Science Symposium (New York City, November 2026). Leaderboard: this https URL ; evaluation harness: this https URL ; public tasks: this https URLSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
We introduce GDP_xlsx, a benchmark of 70 realistic workplace requests written by professionals in 13 domains, each paired with an expert rubric of binary criteria and with the complete workbook or CSV file it refers to. Together the files hold 2.18 million populated cells, and the rubrics 791 criteria. Agents work in a container with spreadsheet tools, and an agentic judge that can open the file applies the rubric. Completeness and exclusivity criteria require set-valued answers to be found in full and with nothing extra. We rank models by strict pass@1, the share of attempts that meet every criterion. Across 20 model configurations, including frontier models from Google, Anthropic, and OpenAI, the top-scoring one, Gemini 4 Argon, passes only 38.3% of attempts, and the median score is 19.3%, leaving wide headroom. We release the evaluation harness and three public tasks; the other 67 are held out to protect the leaderboard from contamination.
- [162] arXiv:2611.17767 [pdf, html, other]
-
Title: Learning Anisotropic Prediction Targets for Flow ModelsComments: Project page: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Flow-based generative models are trained on a hand-chosen prediction target, such as the velocity, the noise, or the clean data. This choice strongly affects training and sample quality. We introduce Anisotropic Flow Models (AnisoFlow), which replaces a fixed prediction target with a fully learnable one. AnisoFlow learns to scale the noise component of the target anisotropically, assigning a separate nonnegative weight to each direction in a jointly learned basis. A closed-form map recovers the velocity from the network's prediction, leaving the architecture, flow-matching loss, and sampler unchanged. Since recovery is exact, training optimizes the target purely to make it easier for the network to predict. We theoretically analyze the resulting prediction error and predict that the preferred noise weight in each direction scales near-linearly with data variance, in agreement with the weights learned empirically in pixel space. AnisoFlow improves FID from 1.90 to 1.55 over the JiT baseline on ImageNet 256$\times$256, by learning only the prediction target, without representation alignment or architectural changes. In pixel-space text-to-image generation, AnisoFlow improves image quality and, through better-conditioned velocity recovery, largely removes patch-seam artifacts under reduced-precision inference.
- [163] arXiv:2611.17768 [pdf, html, other]
-
Title: Activation Steering the Nan-Shan in TyphoonComments: Accepted to NeurIPS Creative AI Track 2026. Website: this https URLSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Sound (cs.SD)
Typhoon is an audiovisual installation based on Joseph Conrad's novella, in which Captain MacWhirr steers the steamer Nan-Shan directly into and through a catastrophic storm. The novella is a study of what happens when the scale of a phenomenon exceeds the practical and conceptual frameworks available to comprehend it and of the paralysis that sometimes follows a demand for action in such overwhelming circumstances. We chose the text of Typhoon to reflect on human culture's current AI moment. We prompt a large language model, seeded with Conrad's original language, to feed endlessly on its own output, and explore when and how the text becomes increasingly alien and unwieldy as time progresses, including deviations from the narrative, digressions into the model's own instructions and thinking, and long periods of incoherent rumination. We employ methods from activation steering, including a novel adaptive steering method we introduce, to control the rhythm of the narrative and the model's digressions over long stretches of generation. The simultaneous efficacy and unpredictability of this steering showcases the tension between the perceived interpretability of AI, and the reality of a system that remains outside our comprehension and control.
- [164] arXiv:2611.17770 [pdf, html, other]
-
Title: Bridge: Database-Grounded Interactive Disambiguation for Text-to-SQLSubjects: Databases (cs.DB)
Ambiguity remains a persistent challenge in text-to-SQL, where an underspecified natural-language question may admit multiple plausible SQL interpretations. Recent work attempts to recover the user's intended interpretation through iterative clarification, but typically relies heavily on LLM parametric knowledge for ad hoc reasoning, leading to limited ambiguity-resolution effectiveness and high LLM cost. We further observe that many ambiguity errors stem from ambiguous column grounding, which remains poorly handled. Therefore, we propose Bridge, a database-grounded disambiguation framework built on an existing iterative protocol. Bridge supplies the LLM's ambiguity detection with offline-distilled knowledge about potential ambiguity relations among columns in the target database. Offline, Bridge constructs a Column Ambiguity Graph by first identifying potentially ambiguous column pairs using complementary database-grounded signals, followed by LLM-based verification. Online, columns and values mentioned in a previous incorrect SQL prediction are used as anchors to retrieve alternatives from the Column Ambiguity Graph and an offline-constructed value similarity index, respectively. The results are integrated for further pruning and consolidated for subsequent ambiguity detection. Extensive experiments on BIRD-dev without the benchmark-provided evidence show that, at comparable LLM cost, Bridge outperforms the strongest baseline by 11.60% in execution accuracy and improves the recall and precision of column usage in the final SQL by 5.25% and 12.37%, respectively.
- [165] arXiv:2611.17772 [pdf, html, other]
-
Title: Antidistillation is Cheap: Lightweight Distillation-Resistance via Stylistic CorruptionComments: Accepted to Findings of EACL 2027Subjects: Machine Learning (cs.LG)
Knowledge distillation (KD) is a highly effective paradigm for transferring capabilities from large language models (LLMs) to smaller, more specialized student models. In commercial deployment where frontier models are restricted behind black-box APIs, sequence-level distillation remains a potent vector for competitors to clone proprietary capabilities at a fraction of the development cost. Existing defenses mostly rely on surrogate models to manipulate next-token sampling, which increases computational overhead and real-time inference latency. In this work, we propose a lightweight defense that dynamically injects style-steering vectors into the teacher model's activations upon malicious queries. This framework offers exceptional flexibility, allowing hosts to customize target defensive styles and dynamically blend them to maximize stylistic variance. Crucially, this inference-time intervention preserves semantic utility and fluency during malicious queries, misleading adversaries into believing that they are extracting pristine teacher outputs, while standard consumers continue to receive original, unmodified responses. Our experiments demonstrate that this framework successfully disrupts student learning, achieving substantial anti-distillation performance while incurring negligible overhead and preserving text naturalness. Our code is available at this https URL.
- [166] arXiv:2611.17773 [pdf, html, other]
-
Title: Compatible by Construction: Bottom-Up Nonlinear Model Reduction without Interface BasesSubjects: Numerical Analysis (math.NA)
Nonlinear manifolds can provide compact representations of solutions to partial differential equations (PDEs), but generating training data and learning these representations on large domains remain costly. We present a bottom-up framework that reuses local nonlinear representations learned on a small training domain to solve larger PDE systems. A shared nonlinear decoder represents the full set of degrees of freedom owned by each subdomain, without separate interior and interface models. Each global degree of freedom has a unique owner, and local residuals directly use decoded state values from the corresponding neighboring subdomains. The resulting discrete state is single-valued by construction, eliminating the need for independent interface copies and additional interface compatibility constraints. Physical coupling is retained through the original time-discrete PDE residual. With the decoder held fixed, the subdomain-specific reduced coordinates are determined jointly by minimizing this residual, and the Gauss--Newton system is assembled from local contributions. Experiments on parameterized viscous Burgers and incompressible Navier--Stokes problems demonstrate reuse on larger domains without retraining. Comparisons with local proper orthogonal decomposition models under the same coupling construction show favorable accuracy--efficiency trade-offs in the tested regimes.
- [167] arXiv:2611.17775 [pdf, html, other]
-
Title: Beyond Prediction Gain: LASR as a Metric for Action-Grounded Latents in World ModelsYu Zhou, Changsong Lei, Yixiang Wei, Kairui Wen, Huankun Sheng, Ming Li, Yeying Fan, Yu-Hui Wen, Yong-Jin LiuSubjects: Robotics (cs.RO); Machine Learning (cs.LG)
World models that support planning, control, and interpretable reasoning require latent representations that genuinely encode action. Yet the large-scale video data from which such models are learned overwhelmingly lack action annotations. Latent-action models infer action-like variables from observation sequences, enabling action-aware learning from unlabeled videos or trajectories. A fundamental concern, however, remains: a latent that improves future prediction may simply encode the current state, episode progress, task identity, or other predictive shortcuts. Existing evaluation practices---reconstruction loss, rollout fidelity, action recovery, or downstream performance---cannot isolate this distinction. We introduce the Latent Action Support Ratio (LASR), a dimensionless metric that measures, under offline evaluation on held-out data, whether a learned latent provides predictive support for action-related targets beyond the current state and non-action alternatives. LASR provides two complementary summaries: GeoLASR captures typical support, while WorstLASR identifies the weakest case. We instantiate the LASR benchmark with recorded actions and corresponding physical state changes as targets, episode-held-out splits, one predictor type for all comparisons, and positive and negative controls. Experiments on recorded robot manipulation data and a matched multi-task panel demonstrate that LASR detects known action information, rejects representative non-action shortcuts, and enables graded comparisons among imperfect latent representations. We further use LASR to train and audit a lightweight residual adapter, obtaining descriptively better point estimates whose uncertainty is assessed with paired comparisons. LASR thus provides a principled offline protocol for measuring action-related predictive support in world-model latents.
- [168] arXiv:2611.17776 [pdf, html, other]
-
Title: Certified Improvements to Best-Known Thomson-Problem Configurations with a Search Pipeline Built with Large Language Model AgentsSubjects: Artificial Intelligence (cs.AI)
The Thomson problem asks for N unit charges on a sphere with minimum Coulomb energy; for large N the best configurations known are curated in the Cambridge Cluster Database, whose largest-N entries were set in October 2026. We report configurations for N = 912, 1382 and 2502 with energies 0.0036186, 0.0109730 and 0.0614033 below the database values, certified by exact integer arithmetic. All three are numerically strict local minima modulo rotations and keep the defect inventory and scar compositions of the database configurations; the database value for N = 912 had been reached by two different methods. Our search runs unconstrained replica-exchange basin hopping from the database configurations, with a shared pool of minima, recombination of improved regions after relabelling by geometry, and gradient-only polishing; replica-exchange basin hopping alone found the N = 912 and 1382 configurations, the former in one of five runs. We built and ran it with large language model agents (Claude, through a chat interface and the Claude Code agent), and we report every search, including those that failed. Two findings matter for others who search the database. Polishing alone lowers two entries by more than the threshold at which our pipeline reports a candidate, which produced two spurious records until candidates were compared with the polished database energies (the polished baseline). A local-relaxation speed-up passed its exactness test yet found nothing in a controlled comparison in which every arm with full relaxation went below the database value within 15 minutes. We distil three practices for computational discovery with large language model agents: calibrate the search code and verify candidates with code separate from the search, compare candidates with the polished baseline, and test design changes in controlled comparisons with pass criteria fixed in advance.
- [169] arXiv:2611.17779 [pdf, html, other]
-
Title: Mobile Edge Deployment and Resource Management for Maritime Wireless NetworksComments: 12 pages, 11 figuresJournal-ref: IEEE Transactions on Vehicular Technology, vol. 74, no. 5, pp. 7928-7939, May 2025Subjects: Networking and Internet Architecture (cs.NI)
Mobile Edge Computing (MEC) has been envisioned as one of the key technologies for supplying computation and storage resources in Internet of Vessels (IoV) networks. Due to its flexible deployment, low cost and agile maneuverability, Unmanned Surface Vehicle (USV) has emerged as a promising solution, to provide communication and computation services for maritime users. In this paper, we study mobile edge deployment and resource management for MEC-assisted maritime wireless networks where USVs with diverse computation resources are deployed to provide edge computing services that complement the cloud-based services. To this end, we formulate an optimization problem to minimize the expected response time by jointly optimizing the deployment of mobile USVs and computation offloading decisions. To solve the mixed-integer nonlinear program problem, we propose a Dual-Layer Reinforcement Learning (DLRL) framework to attain a near-optimal solution. Specifically, a Deep Deterministic Policy Gradient (DDPG) algorithm is designed to obtain the best USV deployment in the outer layer learning, and a Q-learning algorithm is designed to determine the best computation offloading decisions in the inner layer learning. Numerical results demonstrate that the proposed solution outperforms some literature algorithms by effectively handling both continuous and discrete variables.
- [170] arXiv:2611.17780 [pdf, other]
-
Title: LuthorSQL: A One-Pass Text-to-SQL Model as System 1 for a Slower LLM AgentComments: 8-page main text plus appendix, 23 pagesSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Databases (cs.DB); Machine Learning (cs.LG)
Frontier LLM agents that query databases with tools are among the most accurate text-to-SQL systems, but per-query calls to a hosted model are slow, costly, and expose sensitive organizational queries to external APIs. For example, a Claude Opus 5.5 agent serving 10,000 queries a day costs over $70,000 a year at list price. To what extent can local hardware already owned by an organization offload this burden without compromising the agent's accuracy? We present LuthorSQL, a non-autoregressive 7B text-to-SQL model that serves as a rapid System 1 gate for a slower LLM agent (System 2), escalating queries only when uncertain. LuthorSQL makes a single forward pass to predict slot scores over a bounded grammar for query plans, compiles and runs up to 32 candidate plans, and chooses an answer using a trained validator. On its own, LuthorSQL reaches 68.03% execution accuracy on BIRD's hidden test set at 0.84 s per question on a single GPU, maintains its accuracy in 9 GB with 8-bit quantized weights, and transfers zero-shot to Snowflake, BigQuery, MySQL, PostgreSQL, and DuckDB within 3.1 points of its native SQLite accuracy. Combined with Claude Opus 5.5 in a pre-registered confidence cascade over 227 questions on databases published after Opus's training cutoff, LuthorSQL answers nearly half of all questions locally and cuts API cost by 33.3% at the same accuracy. Finally, contamination probing shows that Opus reproduces 48-92% of public benchmark gold queries verbatim without seeing the schema; however, its tool-assisted accuracy holds on renamed schemas and reaches 95.6-97.4% on post-cutoff databases. Code and data: this https URL
- [171] arXiv:2611.17781 [pdf, html, other]
-
Title: TReC: Evaluating Move-Induced State Changes in Instruction-Conditioned GoSubjects: Artificial Intelligence (cs.AI)
When a Go model predicts a move different from the expert's, coordinate distance measures how far the prediction lands from the reference. We introduce Transition Representation Consistency (TReC) to compare how the two moves change a board-state embedding. Both moves are executed from the same position under the game rules, and TReC is the cosine similarity between the two resulting representation changes. In experiments with a frozen KataGo network on 680 mid-to-late-game positions from five games, the gap between the highest and lowest candidate per position has a median of 0.01 to 0.08 for the direct successor-state cosine and 0.88 to 1.33 for TReC, depending on the evaluator layer. TReC follows coordinate distance only weakly: the rank correlation between a candidate's distance to the reference and its TReC, computed within each position, has a median of -0.19 over these positions.
- [172] arXiv:2611.17784 [pdf, html, other]
-
Title: Direction or Magnitude? Learning Scalars over Task-Calibrated DirectionsSubjects: Machine Learning (cs.LG)
Parameter-efficient fine-tuning (PEFT) aims to preserve adaptation quality while reducing training costs. We propose SCALAR, which achieves performance comparable to LoRA with substantially fewer trainable parameters, lower peak training memory, and higher throughput. SCALAR constructs a fixed input basis from task activations and assigns a random output dictionary to each input direction. Fine-tuning learns only layer-specific scalar coefficients, reducing gradient and optimizer-state storage. Grouped computation aggregates output atoms before applying the adapter to token activations, while dictionary sharing across compatible layers reduces fixed-basis storage. Experiments on commonsense reasoning, mathematical reasoning, code generation, and knowledge evaluation after instruction tuning demonstrate effective adaptation at small trainable budgets. On commonsense reasoning, SCALAR achieves accuracy comparable to LoRA with about 4% of its trainable parameters. System measurements on Qwen2.5-7B show 20% lower peak training memory and 6% higher throughput than LoRA. The combination of adaptation quality, a compact trainable state, and training efficiency makes SCALAR a promising approach to resource-constrained fine-tuning.
- [173] arXiv:2611.17785 [pdf, html, other]
-
Title: CHARMING: Parallel Edge AI Acceleration Across Tensor Processing and In-Memory ComputationMd Omar Faruque, Jingang Jin, João Paulo C. de Lima, Shouzhi Fang, William C. Tegge, Peipei Zhou, Qinru Qiu, Alex K. JonesSubjects: Hardware Architecture (cs.AR)
Edge AI increasingly relies on specialized accelerators for low-latency inference under tight resource and energy constraints, yet their efficiency remains limited by data movement. Prior work, CHARM, mitigates this overhead by overlapping data movement with computation. We observe that in-memory computing can further supplement tensor processing as a parallel accelerator while eliminating certain data movements entirely.
We present CHARMING, a heterogeneous edge AI architecture that uses accuracy- and quantization-aware partitioning to distribute and concurrently execute inference workloads across tensor processing and in-memory computation. CHARMING partitions attention heads between an AIE array and an in-DRAM compute unit, balancing execution according to model architecture, sequence length, and precision while selecting heads to minimize accuracy degradation. CHARMING also integrates quantization into the AIE projection stage to further reduce data movement. We validate CHARMING end-to-end in hardware on AMD Versal VCK5000 and Alveo U200 platforms and evaluate LLM and video-generation models across multiple precisions using measured hardware results and simulation of the latest in-memory compute hardware.
CHARMING improves prefill performance by up to 23.6 percent and decode performance by up to 44 percent over AIE-only acceleration. With int4 x int4 fixed-point in-memory computation, CHARMING reduces per-layer energy by up to 7.6 percent and improves energy-delay product by up to 24.7 percent, while maintaining accuracy similar to a 32-bit floating-point baseline. - [174] arXiv:2611.17786 [pdf, html, other]
-
Title: Towards Seconds-level Video Representation: Quadratically Gated Feature Grid with Adaptive, Prediction-free SamplingTie-Yun Li, Hao-Ran Wang, Chao Wang, Zhiyong Yin, Yiming Xu, Zhenwei Cai, Li Li, Rongzhi Lin, Feng-Lei FanComments: 13 pages, 11 figures, 4 tables. Code available at this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Implicit neural video representations are notoriously slow to build, which does not cross the basic practical threshold for most industrial applications. Here, we present Q-GAS, a coordinate-queryable video representation that couples multi-resolution explicit 3D feature grids with a compact SwiGLU-style quadratic-gating (QG) decoder and an error-aware, prediction-free sampler. The explicit grids avoid hash-collision coefficient sharing at the cost of modestly higher storage yet greatly boost the expressive capacity; the QG decoder introduces low-rank local second-order interactions to improve the nonlinear feature response. During fitting, a global error map periodically refreshes a high-error pool, and each batch mixes pool samples with uniform samples, which is free from the auxiliary model to predict difficult samples, thereby further escalating the efficiency. On UVG Beauty/Bosphorus/Jockey, Q-GAS fits and quantizes under 50 s on one RTX PRO 6000, decodes at over 370 FPS with mixed precision and about 146 FPS in FP32. In our PyTorch implementation, Q-GAS trains and quantizes 5.5-7.2 times faster than the evaluated Instant-NGP baseline and 19-28 times faster than RNeRV. Under a 90 s budget, it improves PSNR/MS-SSIM/LPIPS over the evaluated NeRV-family pipelines at comparable BPP over the measured overlap. We have shared our code at this https URL for readers' free use.
- [175] arXiv:2611.17788 [pdf, html, other]
-
Title: RAG-FGO: Dynamic Optimization of Query-Rewriting Agents for Generative RAGQigan Sun, Xudong Wang, Chaoning Zhang, Chenghao Li, Jiaquan Zhang, Jiehui Xie, Sungyoung Lee, Jie Zou, Ning Xie, Peng Wang, Yang Yang, Heng Tao ShenComments: 12 pages, 6 figures, 8 tablesSubjects: Computation and Language (cs.CL)
In retrieval-augmented generation (RAG), query rewriting serves as a core interface that aligns user intent with external knowledge bases, and its effectiveness directly determines overall system performance. However, generative retrieval faces three critical challenges in real-world deployment: noisy retrieval feedback, the difficulty of jointly optimizing discrete prompt templates and continuous decoding-control parameters, and the limited generalization of static rewriting strategies under semantically complex scenarios. To address these issues, we propose RAG-FGO, a dynamic optimization framework for generative retrieval strategies. RAG-FGO formulates query rewriting as a non-differentiable black-box optimization problem in a mixed discrete-continuous space, and represents prompt templates and decoding parameters as adaptively optimizable query-rewriting agents. The framework performs global search to enable semantic jumps and structural reorganization that escape local optima, while applying local perturbations for fine-grained alignment and precise refinement. In addition, RAG-FGO introduces an elite-preserving memory pool that accumulates high-quality strategies across iterations, thereby mitigating the impact of stochasticity and noisy feedback on optimization stability. Experimental results on a diverse suite of benchmarks demonstrate the effectiveness and robustness of the proposed method, supporting its applicability to complex semantic retrieval tasks.
- [176] arXiv:2611.17789 [pdf, html, other]
-
Title: Can Large Language Models Optimize Pretrained Neural Network Weights? A Multi-Phase Empirical Evaluation of Gradient-Free Editing in YOLO26 Object DetectionSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Direct weight editing offers a gradient-free alternative to fine-tuning pretrained detectors, but its generalization and the value of large language model (LLM) guidance remain uncertain. We evaluate five experimental phases using YOLO26 models and MS COCO, comparing deterministic scaling, random perturbations, differential evolution, CMA-ES, structured interventions, real local-LLM proposals, failure-aware guidance, and hybrid LLM-CMA-ES against a matched random-policy control. Optimization uses native Ultralytics mAP50-95 on 1,000 COCO Val2017 images; selected candidates are assessed on a disjoint, previously examined 1,500-image subset, with COCO API AP50-95 reported separately where available. Increasing the CMA-ES budget from 8 to 250 evaluations raised development gains from +0.11 to +0.28 percentage points, while selection changes remained negative (-0.03 to -0.12 points); the 250-evaluation result uses one seed. The hybrid LLM approach showed no statistically established advantage over random-policy guidance (p = 0.612, 0.369, and 0.694 at budgets 8, 25, and 50). Five-seed YOLO26m replication yielded a COCO API AP difference of +0.004 points (95% CI: -0.026 to +0.034). Of 225 archived LLM proposals, 53 satisfy the final strict schema; early tensor-level execution cannot be fully verified. Genuine LLM-mediated weight editing is feasible, but the tested approaches did not demonstrate consistent transfer beyond the development subset or a measurable LLM advantage.
- [177] arXiv:2611.17790 [pdf, html, other]
-
Title: AutoRL: A Multi-Agent System towards Autonomous Reinforcement LearningSimon Tang, Wenjie Shi, Yongjing Yin, Keheng Wang, Jialin Liu, Zheng Zhang, Linvo Lin, Xiangyuan Xue, Kan Ren, Jingang Wang, Xunliang CaiComments: 23 pagesSubjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Recursive self-improvement (RSI) is one of the most promising paths toward an AI singularity for superintelligence. Achieving this vision requires AI systems to undertake the full training pipeline for iterative improvement. Reinforcement learning (RL) plays an essential role in this cycle to improve model capabilities and generalization, yet automating RL remains difficult due to unstable training dynamics and the need for sustained supervision. We introduce AutoRL, a multi-agent system for automating RL training, as a cornerstone for future RSI. Different from conducting research by end-to-end complete experiments, AutoRL monitors the evolving training state and integrates process-level supervision and research into an ongoing RL training, following the practical workflow of an RL researcher. Specifically, AutoRL consists of three RL-specific agentic roles. Perceptors track training dynamics, rollouts, evaluation, and efficiency, triggering deliberation when they detect anomalies, degradation, or stagnation. Three complementary Proposers diagnose problems and refine candidate interventions through shared evidence and checks. The Orchestrator reviews their recommendation and applies approved changes through online hot updates or checkpoint-based resumption. Experiments on Qwen3 and LongCat-Flash-Lite demonstrate the superiority of AutoRL over static RL recipes and a general agent. The research records show that existing agents can handle unattended RL supervision and dynamic adjustment. The difficulty for future RSI lies in developing fundamental methods that are valuable for resource-consuming validation.
- [178] arXiv:2611.17793 [pdf, html, other]
-
Title: MASTARS: Multi-Agent Sequential Trajectory Augmentation with Return-Conditioned SubgoalsSubjects: Artificial Intelligence (cs.AI)
The performance of offline Multi-Agent Reinforcement Learning (MARL) is limited by the quality of the fixed offline dataset, motivating trajectory augmentation. However, augmenting multi-agent trajectories presents a fundamental challenge of the diversity-coordination trade-off: independent generation improves diversity but lacks coordination, while joint generation induces coordination but lacks diversity, reproducing the training distribution. To address this challenge in multi-agent data augmentation, we introduce MASTARS, a novel diffusion-based framework that generates multi-agent trajectories sequentially across agents while enforcing cross-agent consistent coordination via the method of inpainting. MASTARS produces trajectories that are diverse, coordinated and globally coherent. Experiments on various benchmarks such as MPE, SMAC, and SMAC-v2 demonstrate significant performance gains across multiple offline MARL methods.
- [179] arXiv:2611.17794 [pdf, html, other]
-
Title: Federated Physics-Informed Thermal Field Prediction Across Unseen Geometries in Distributed Additive ManufacturingSubjects: Machine Learning (cs.LG)
In additive manufacturing (AM), each change in part geometry requires a new thermal simulation, which makes design evaluation over many candidates costly. Geometry-aware models reduce this cost by predicting thermal fields on unseen parts with a single learned model, but they assume that diverse geometries are available centrally. This assumption is restrictive when each manufacturing site holds only a limited set of proprietary part designs that it cannot share. We propose Fed-PIGANO, a federated framework that trains physics-informed geometry-aware neural operator (PI-GANO) models at individual sites and aggregates their geometry encoders and prediction networks without sharing part geometries. We evaluate the framework on a 2D steady-state thermal benchmark for metal AM, with geometries from five polygonal families distributed across three sites under four scenarios that range from identical to strongly different family compositions. Over seven replications, Fed-PIGANO achieves mean relative L2 errors of 1.83% to 2.04% across all 12 scenario-site cases, compared with up to 20.46% for local-only training, and consistently yields lower PDE residuals. Its prediction accuracy is comparable to that of a centralized model trained on the pooled geometries of all sites. These results demonstrate that federated learning lets manufacturing sites benefit from each other's geometries without sharing them, enabling rapid thermal prediction on new part designs for tasks such as shape optimization, inverse design, and uncertainty quantification.
- [180] arXiv:2611.17796 [pdf, other]
-
Title: Robustness Evaluation of GPU Cloud Pricing Benchmarks under Incomplete Observations: A Controlled Missingness StudyComments: 5 pages, 3 figures, 7 tablesSubjects: Artificial Intelligence (cs.AI)
Daily GPU rental-price benchmarks can fail when marketplace quotes are missing. This article stress-tests a fixed equal-weight median rule under four imposed masking mechanisms and four benchmark-level recovery methods. Across 5,800 Monte Carlo realisations in 27 scenarios on 190 held-out days, whole-source outages reduce publication coverage to 0.77 while preserving zero conditional error on published dates; composition-changing loss retains 0.86 coverage but produces 7.74 log points of error. A separate past-only stratum-level ablation reduces error by 4.15 to 6.70 log points in five composition-changing cases, using a richer information set. An observable-signal withholding rule improves the primary error-coverage criterion in 0 of 10 cases. Results concern observed asking-price measurements, not transaction prices or identified real-world missingness mechanisms.
- [181] arXiv:2611.17797 [pdf, html, other]
-
Title: Vision-as-a-Tool: Visual Harness for Multimodal Agentic CodingHaolin Yang, Wenshan Wu, Jin Jiang, Yu Cheng, Yan Xia, Ting Song, Yangyu Huang, Zhiliang Peng, Wenhui Wang, Hangbo Bao, Yuzhe Zhang, Wen-Da Wei, Tengchao Lv, Li Dong, Lei Cui, Furu WeiComments: 27 pages, including references and appendices. Project page: this https URLSubjects: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
Multimodal agentic coding requires robust visual reasoning to analyze observations and evaluate generated artifacts. Current frameworks largely rely on a Thinking-with-Images paradigm, where agents depend on their intrinsic vision to execute visually grounded tasks. However, this architectural coupling introduces a fundamental bottleneck. Agent performance remains strictly bounded by internal perceptual limitations. Furthermore, text-only agents lack native perception. Without specialized tools, they cannot perform visual reasoning. To overcome these constraints, we introduce the Vision-as-a-Tool (VaaT) paradigm. Prior visual tool-use merely manipulates images for an agent's intrinsic vision to read. In contrast, VaaT reframes general-purpose multimodal large language models (MLLMs) as callable external utilities. As a result, text-only agents can completely offload visual reasoning during agentic execution. Meanwhile, vision-capable agents can dynamically augment their native perception on demand. We instantiate this paradigm in VHarness, a unified framework driven by five atomic actions. We comprehensively evaluate VHarness across Image-to-HTML, Image-to-SVG, and Image-to-3D generation tasks. Experimental results demonstrate that VHarness consistently improves visual fidelity and geometric accuracy across both text-only and multimodal coding agents. These findings highlight that modularizing visual reasoning into callable utilities provides a flexible and scalable foundation for future agentic coding systems.
- [182] arXiv:2611.17798 [pdf, html, other]
-
Title: Elasticy: A Runtime-Reconfigurable Architecture for Dynamic StreamsComments: 14 pages, 11 figures, 7 tablesSubjects: Hardware Architecture (cs.AR)
Reconfigurable Dataflow Accelerators (RDAs) have demonstrated high performance and energy efficiency for dense, regular tensor applications. However, conventional RDAs with static configurations show dramatically reduced performance and efficiency on irregular tensor workloads with data-dependent, non-uniform, dynamic tensor shapes. We propose Elasticy, a dynamically reconfigurable dataflow accelerator that employs streaming and tile abstractions to minimize the overhead of dynamic support. By encoding dynamic shapes into streams using rank-variable expressions, Elasticy exploits data-dependent tensor shape dynamism to minimize off-chip data traffic and wasted compute cycles. Elasticy employs stream-embedded reconfiguration information to trigger fine-grained runtime scheduling decisions. Elasticy also introduces a novel distributed linked streaming memory (DLSM) that more effectively uses on-chip memory to minimize off-chip data traffic and support dynamically sized tensors.
Both reconfiguration and shape dynamism incur minimal overhead in Elasticy: reconfiguration occurs infrequently at the granularity of a stream of tiles, while rank-variable expressions enable lightweight handling of dynamic dimensions. On a Llama 3-8B decode layer, Elasticy achieves 92.6% of its roofline performance, compared to 69.7% on an H100. On evaluated GEMMs, the DLSM reduces total off-chip traffic by up to 61.4% relative to the best evaluated cache configuration, with no additional burden of programmer-managed memory. Elasticy's distributed runtime scheduler overlaps dependent kernels, speeding up kernel pairs by up to 1.42x. Programs can also run in both space and time: each phase of a layer's graph is pipelined spatially, while phases fold onto the same fabric over time. Compared to a traditional spatial accelerator, Elasticy achieves a 6.34x-31.67x improved area delay product on full layers. - [183] arXiv:2611.17802 [pdf, html, other]
-
Title: Multimodal Person Concept Model for Person Identification under Varying Observation ConditionsComments: 12 pages, 4 figures, 11 tables. Submitted to Artificial Life and Robotics (AROB Journal Track). Project page: this https URLSubjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
Service robots in domestic and office environments must identify users to provide personalized services based on their preferences and previous interactions. This is challenging because only a few observations may be available for learning, while observation conditions change owing to appearance, environmental noise, and utterance content. To address this challenge, this study proposes a multimodal person concept model that learns person concepts from a small number of observations, without person labels, and enables robust identification under changing conditions. The model employs a nonparametric Bayesian generative model based on the Chinese restaurant process and integrates visual, acoustic, and linguistic features extracted using Contrastive Language-Image Pre-training (CLIP) and Contrastive Language-Audio Pretraining (CLAP), followed by principal component analysis (PCA). We evaluated six participants under eight conditions in which appearance changes, additive noise, and transcription errors were introduced individually and in combination. Averaged over the eight conditions, the proposed model achieved an identification accuracy of 0.89 while inferring the number of person concepts from the observations, which is comparable to k-means clustering that was supplied with the true number of participants (0.88). We also examined PCA dimensionality, modality availability, and variation strength, and demonstrated the model on a Toyota Human Support Robot. These results show that multimodal person concept learning supports robust person identification under changing conditions, and that identification can succeed even when the number of concepts is not correctly recovered. The project page and demonstration video are available at this https URL.
- [184] arXiv:2611.17803 [pdf, html, other]
-
Title: Resource Management and Trajectory Optimization for UAV-IRS Assisted Maritime Edge Computing NetworksComments: 17 pages, 9 figuresJournal-ref: Tsinghua Science and Technology, 2025, 30(4): 1600-1616Subjects: Networking and Internet Architecture (cs.NI)
With the exponential growth of maritime wireless devices and the rapid development of maritime applications, traditional maritime communication networks encounter communication and computation limitations in supporting computation-intensive and latency-critical tasks. Edge computing and Intelligent Reflecting Surface (IRS) have emerged as promising techniques to improve communication and computation services for maritime devices with limited computation capabilities and battery capacity. This paper studies an IRS Mounted on Unmanned Aerial Vehicle (UIRS) assisted maritime edge computing network, in which the UIRS is deployed to assist the transmission from Unmanned Surface Vehicles (USVs) to the edge server via Non-Orthogonal Multiple Access (NOMA) protocol. We propose a resource management and trajectory optimization scheme by jointly optimizing subslot duration, offloading ratios, transmit power, edge computation capability allocation, UIRS phase shifts and UIRS trajectory, aiming at minimizing the overall energy consumption. Since the non-convex nature of the optimization problem, we propose a two-layered method by decomposing the original problem into two subproblems. The top-layered subproblem is solved by the Semi-Definite Relaxation (SDR) method and the underlying-layered subproblem is solved by the Deep Deterministic Policy Gradient (DDPG) algorithm. Numerical results demonstrate that our proposed scheme can effectively and efficiently reduce overall energy consumption.
- [185] arXiv:2611.17806 [pdf, html, other]
-
Title: Can Agents Keep Up with Time? An Empirical Analysis of Temporal Adaptation After a Conversational HiatusSubjects: Artificial Intelligence (cs.AI)
Long-horizon agents increasingly serve as persistent personal assistants across modern web platforms, managing schedules, applications, and multi-step plans over weeks or months. Users often return to an ongoing conversation after a long conversational hiatus, while earlier bookings and preparation notes stay on record. During this hiatus, scheduled dates may pass even though the retained records remain unchanged and event outcomes go unrecorded. Existing evaluations of long-horizon agents focus on fact retention and explicit state updates. Whether agents perceive the passage of time and adapt their advice accordingly therefore remains unknown. We frame this ability as temporal state perception and use, which means perceiving the current time and applying it when interpreting dated records. To study it, we design session-rekindling tasks that keep the earlier conversation fixed and vary only the follow-up date and the way temporal information reaches the agent. Experiments across five model routes and five temporal-information conditions show that agents can obtain the current time but often fail to use it. They reliably query the clock and even note in their reasoning that the scheduled date has passed, yet many still tell users to prepare for the expired event. Supplying explicit temporal context reduces this error, and providing the computed date relation removes it entirely. Keeping up with time therefore depends on turning perceived time into updated plans. Based on these findings, we design a session-rekindling benchmark that provides temporal-awareness guidance for designing long-horizon agents. The code can be found at this https URL.
- [186] arXiv:2611.17808 [pdf, html, other]
-
Title: CarryOn: Learning What to Keep in Context for Long-Horizon LLM AgentsSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Long-horizon LLM agents can coordinate tools and applications to complete complex, multi-step tasks. As their interaction histories grow, effective context management becomes essential for reliable and efficient execution. Existing methods reduce context through summarization, pruning, or history selection. However, shorter contexts do not necessarily support better decisions: summaries can omit exact details needed for later actions, while retaining entire interactions can carry substantial irrelevant information. We propose CarryOn, a lightweight context manager that pairs compact execution state with fine-grained, source-linked evidence. A learned selector and adaptive retention policy keep relevant evidence available without additional LLM calls for context management. We introduce two complementary evaluation metrics: $S_\alpha$ prioritizes task success while accounting for resource use, and normalized cost per success (NCS) measures the resources spent per successful task. Across AppWorld, OfficeBench, and Tau2 Telecom with three LLM Actors, CarryOn achieves 75.7% macro-averaged task success, exceeding the strongest baseline by 6.7 percentage points. It also achieves the best average scores on both metrics.
- [187] arXiv:2611.17811 [pdf, html, other]
-
Title: AI Reviews Are Biased Towards the AppendixSubjects: Computation and Language (cs.CL)
Large language models (LLMs) are increasingly used by reviewers to highlight weaknesses during the reviewing process, and their outputs often shape the reviews that humans submit. Prior work studies how LLM reviews align with human ones, yet we know little about which parts of a paper these models actually criticize. In this work, we investigate 2,000 papers from ACL, EMNLP, COLM, and ICLR, and show that the latest GPT and Claude models draw about 50% or more of their review weaknesses from the appendix. This behavior holds across the API, coding-agent, and web access methods for using LLMs to review a paper. In contrast, we find that humans draw only 8.4% of their weaknesses from the appendix, and these weaknesses primarily revolve around a paper's presentation and content related to its core idea. Meanwhile, LLMs draw weaknesses from deep in the appendix that concern fine-grained technical details not central to the paper's contributions. We anticipate that our analysis will contribute to the scholarly discourse on AI reviews. Our code is publicly available at: this https URL.
- [188] arXiv:2611.17814 [pdf, html, other]
-
Title: G-CAST: Governed Continual Self-Training for E-Commerce Agents under Evolving DemandsSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
With the rapid advancement of Large Language Models (LLMs), an increasing number of users are turning to LLM-powered agents for conversational shopping in e-commerce. However, the user needs continue to change as new product categories, procurement scenarios, and feature requests emerge. At the same time, the product's execution harness-including its tools, skills, and operating procedures-continues to evolve to support increasingly sophisticated interactions. These changes create two practical challenges: 1) Demand drift alters the distribution, intent, and complexity of user requests, causing a model trained once to fall behind; 2) Harness evolution requires the model to use newly introduced capabilities correctly without losing behavior that users already rely on. Periodic manual retraining is too delayed and labor-intensive for this setting. To address these challenges, we present G-CAST, a governed continual self-training system that converts daily production evidence under evolving demand and capability contracts into reward-routed, entropy-prioritized training data and governed reward evolution. Governed checkpoint transition under cumulative evaluation controls each model update, enabling the agent to learn newly introduced behaviors without sacrificing previously accepted capabilities. Experiments on a real production e-commerce agent show that G-CAST achieves stronger adaptation to newly introduced demand while better preserving previously accepted behavior compared with a strong autonomous post-training baseline. Ablations further show that both reward-routed, entropy-prioritized data orchestration and governed reward evolution are important to these gains. G-CAST is now deployed in Yuancai(this http URL), a B2B procurement agent on Alibaba's 1688 platform.
- [189] arXiv:2611.17815 [pdf, html, other]
-
Title: Improved Closed-Form Identification of the Air-Gap Permeance in Magnetic Equivalent Circuits of Rotor-Eccentric Electric MotorsComments: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessibleSubjects: Systems and Control (eess.SY)
Fast and accurate simulation of electric motors is essential for their design, fault diagnosis and control. Magnetic equivalent circuits (MECs) offer a computationally efficient alternative to the finite element method (FEM), yet their flux tube representation poorly captures the air-gap field. This creates a challenge, as the air gap stores most of the magnetic energy and governs the electromagnetic torque and, in rotor-eccentric motors, the unbalanced magnetic pull (UMP). Existing air-gap permeance expressions vary widely in their results for torque ripple and UMP and account for tooth-normal eccentricity only. This paper proposes the dual-term air-gap permeance (DTAP) model, a closed-form expression for the air-gap permeance between a stator and a rotor tooth that accounts for tooth-normal and tooth-tangential eccentricity, with and without rotor skew. Its parameters are identified from a FEM model of the air gap and the complete set of tooth heads, which includes the shielding by the neighbouring teeth. For three motors, DTAP reproduces the FEM air-gap permeance with a 2.8-4.8 times smaller error than the state of the art, at a comparable computational cost. Implemented in the MEC of an eccentric induction motor, it reproduces the torque ripple within 1.6% and the peak UMP within 1.3% of the reference, i.e., with 2.8 and 27 times smaller deviations than the state of the art. DTAP thereby enables MEC to predict the torque pulsations and bearing loads of eccentric motors more accurately without sacrificing computational speed.
- [190] arXiv:2611.17817 [pdf, html, other]
-
Title: Quantum Shock and Train-Time Constrained Quantization: A Geometric Study of Finite-Resolution Variational Quantum CircuitsSubjects: Networking and Internet Architecture (cs.NI)
Finite-resolution interfaces map Variational-Circuit angle to discrete codebook. Since Pauli rotations are periodic, so Euclidean rounding may produce wrong codeword near a chart boundary; error due to coordinates alone cannot measure how much the final state was changed. Herein we define a metric, Quantum Shock, for Pure State it is defined as the average of the square of the Fubini-Study distance (over all inputs) between the pre-deployment and deployed states. This value is limited by constant multipliers of the mean of infidelity and provides upper limits on displacement when observing bounded quantities and margin based predictive stability. Extensions to the analytical Bures-angle for mixed-states were derived, along with a finite shot stability bound. The proposed Train-Time Constrained Quantization allows a delay in applying a penalty for selecting a codeword aligned with periodicity until after task only optimizations. Three small binary tasks were evaluated on a four qubit, two layer classifier using ten matched seeds, K = 4, 8, 16 and exact expectation values from simulation under noiseless conditions and state vector simulations. On our primary evaluation, T-CQ exhibited smaller average local quantum shock compared to circular post-training quantization for each task and code-book size tested. The re-execution on the x86-64 platform using K=8 preserved the direction of the mean-shock comparison but had varying magnitudes.
- [191] arXiv:2611.17818 [pdf, html, other]
-
Title: LLMs are Vulnerable to Malicious Requests with Indistinguishable IntentSubjects: Cryptography and Security (cs.CR)
Jailbreak attacks can circumvent the safety mechanisms of Large Language Models (LLMs) and lead to the generation of harmful or policy-violating content. We first examine eight jailbreak methods against an LLM-based filter and observe that requests resembling normal user queries tend to have higher filter pass rates than those containing obvious anomalous patterns. This observation motivates the hypothesis that resemblance to normal requests may contribute to jailbreak success. We introduce \textbf{Detail Guidance Attack (DGA)}, a framework that uses an auxiliary LLM to construct detailed requests and interacts with the victim model in a single turn. Across 11 victim models and three datasets, the complete DGA pipeline achieves high attack success rates under our automated judging criterion, including rates above 95\% in several evaluated settings. We also evaluate a DGA-aware filter and conduct a user study on everyday requests with ambiguous or dual-use intent. The survey documents differing expectations about whether models should respond; the usability costs of the filter remain to be measured.
- [192] arXiv:2611.17819 [pdf, html, other]
-
Title: Causally Constrained and Uncertainty-Aware Cardiopulmonary Risk PredictionSubjects: Artificial Intelligence (cs.AI)
Routine chest computed tomography (CT) captures pulmonary, cardiac, and vascular phenotypes that jointly stratify lung cancer and cardiovascular disease, the two leading causes of death in lung-screening populations. Yet clinical risk calculators evaluate each disease in isolation and omit CT findings, whereas unconstrained machine learning overfits small hospital cohorts, learns treatment- and illness-confounded risk directions, and conceals uncertainty when patient records are incomplete. Here we present the Causal Envelope Agent (CE-Agent), a two-stage framework that compiles peer-reviewed epidemiological evidence into explicit constraints on risk-factor weights before any outcome is seen, while estimating local baseline risk and newly added CT phenotypes from cohort data. At deployment, closed-form uncertainty propagation accounts for missing clinical or CT inputs within milliseconds per patient, running over two orders of magnitude faster than Monte Carlo sampling and returning calibrated risk intervals that direct chart review and selective imaging to borderline patients. Across 53,452 lung-screening participants from 33 US hospitals and 6,814 multi-ethnic adults, CE-Agent outperforms cardiovascular calculators in lung screening, exceeds the Pooled Cohort Equations on hard atherosclerotic events while correcting their 1.7-fold risk overestimation, and generalizes from a single hospital or CT scanner to external sites while lowering required sample sizes up to fivefold. Applied unchanged to lung cancer, CE-Agent matches established screening models with accurate calibration even when restricted to risk-factor directions alone, and pairing both predictions improves joint cardiopulmonary triage from a single screening visit.
- [193] arXiv:2611.17820 [pdf, html, other]
-
Title: Aligning AI TeamsSubjects: Artificial Intelligence (cs.AI)
Previous work has shown that alignment training on individual LLM agents can fail to transfer to multi-agent settings. We investigate the underlying mechanisms for this failure, and explore interventions to preserve alignment transfer to multi-agent LLM teams. We focus on two software-engineering tasks (sepsis triage, a news recommender) and ten consulting-proposal generation tasks. We present four findings: (1) The single-vs-team safety gap does not show a clear trend with more capable models. We find that Anthropic's Mythos Preview shows the largest gap on one of our tasks but the smallest on another; (2) Many natural interventions, such as having agents critique each other's work or coordinate in a group chat, do not reliably bring agent teams back to single-agent alignment behavior; (3) diffusion of responsibility substantially accounts for the safety gap: agents in teams are less likely to proactively check for system-level safety failures and more likely to ignore, rationalize, or deflect responsibility for issues when they arise; (4) having a designated safety lead agent ensure system-level safety, and having a coordinator agent publish a safety specifications document are effective at restoring single-agent alignment. We also release magelab, a multi-agent LLM orchestration and experimentation framework. Our findings can help practitioners design agent teams that successfully preserve the behaviors expected of aligned individual agents.
- [194] arXiv:2611.17821 [pdf, html, other]
-
Title: SceneAgent: A Multi-Agent Framework for Portable Game Scene GenerationComments: Project page: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Moving a scene generator to a new game engine requires more than exporting its final output: planning, construction tools, and stored experience may also depend on the original backend. We present SceneAgent, a multi-agent framework for reusing the construction process and its experience across backends. A canonical SceneContext maintains the evolving world, while backend adapters isolate coordinate conventions, asset bindings, and export. Structured handoffs let fresh Builder sessions query this persistent state without inheriting the full construction transcript. Across generations, human-approved traces are distilled into asset-ID-free Layout skills and semantic Content skills that can be retrieved and recombined. We introduce GameSceneBench, comprising 69 indoor and outdoor prompts. Under a common asset and evaluation protocol, SceneAgent obtains human tie-adjusted preference scores of 78.2-89.4% against five baselines. Ablations show that reusable skills and the multi-agent protocol primarily improve prompt alignment and scene completeness, with additional runtime. Deployment in a second engine preserves the shared construction protocol and skill library while replacing backend integration components. Together, these results support separating persistent scene state, task-local reasoning, and reusable experience for portable scene construction.
- [195] arXiv:2611.17822 [pdf, html, other]
-
Title: Scalar Entropic Damage Transport: Well-posed softening with a single damage variableComments: 42 pages, 15 figures. Submitted to Computer Methods in Applied Mechanics and EngineeringSubjects: Computational Engineering, Finance, and Science (cs.CE); Numerical Analysis (math.NA)
Entropic Damage Transport (EDT) represents material softening via a local continuum damage model while ensuring the well-posedness of the mechanical boundary value problem. It is based on a convexification of the time-incremental potential associated with a rate-limited scalar damage model. A one-dimensional distribution of damage states serves as the internal variable, and the number of these states grows with the number of time steps. We show that the emergence of the damage distribution is an artifact of the time discretization. Surprisingly, the distribution concentrates at a single damage state in the time-continuous limit, and EDT converges to a rate-limited damage model with a scalar damage variable, whose free energy is jointly convex in the strain and the reciprocal degradation. We exploit this hidden convexity by a frozen-metric discretization in terms of the reciprocal degradation, which renders every increment jointly convex. The resulting scalar update preserves damage irreversibility, the prescribed damage-rate limitation and thermomechanical consistency, reduces to a scalar equation and provides a symmetric and positive definite algorithmic tangent in closed form. Computational experiments assess the accuracy and the efficiency of both formulations from the material point to the microstructure level, and demonstrate that the scalar update permits a finite-element computation of industrial size with more than 14 million degrees of freedom.
- [196] arXiv:2611.17824 [pdf, html, other]
-
Title: TailGuard-GDS: Tail-Safe Control of GPU Storage Data MovementComments: This is a full-length research article. The manuscript is 35 pages and contains 14 figures and 17 tables, including the appendix. It is not under consideration elsewhereSubjects: Distributed, Parallel, and Cluster Computing (cs.DC); Operating Systems (cs.OS)
AI runtimes move state between NVMe storage and GPU memory on critical paths. GPUDirect Storage (GDS) removes host staging, yet our eight-H100 measurements show it is not consistently tail-optimal for semantically identical GPU-buffer operations. The preferred path varies with operation, request size, concurrency, placement, and resource contention; bandwidth and tail-latency winners can disagree. Frameworks specify what moves and when, while GPU I/O stacks expose mechanisms; neither combines phase semantics with measured tail behavior to choose a delivery plan.
We present TailGuard-GDS, a runtime that selects a semantically equivalent plan per application phase. It profiles plans varying backend, request shape, and placement, minimizes GPU-visible tail latency under phase constraints, and revisits decisions as conditions change. Resource reuse keeps policy lookup and switching at microsecond-scale overhead.
Integrated with PyTorch FSDP and DeepSpeed ZeRO-3, TailGuard reduces FSDP step p95 by 18.4% and GPU-idle p95 by 25.3% versus the strongest fixed path. Across five eight-GPU ZeRO-3 workloads, step p95 falls 7.9-24.6% versus fixed host staging and 28.4-42.6% versus fixed GDS. Against tuned DeepSpeed AIO, TailGuard reduces step p95 by 5.6% and raises estimated throughput by 6.1%. - [197] arXiv:2611.17825 [pdf, html, other]
-
Title: Beyond Endpoints: Bézier Paths for Vision Foundation Model AdaptationComments: 47 pagesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Frozen vision foundation models can be adapted by learning a few prompt tokens shared across all images. When prompts are updated as data shift or tasks change, the distribution of image features moves from one endpoint to another. Prior theory asks which endpoints prompts can reach; we ask whether they can travel between them along the shortest path. Because shared prompts compete with each image's own content for attention, they often cannot. We quantify this restriction as the extra transport cost that shared attention imposes over the shortest path. In an exactly solvable attention model, the reachable features form a cone; unrolling it into a plane gives the exact minimum action for every pair of endpoints and explicit optimal prompts on short arcs. The analysis reveals a sharp threshold in content normalization, which measures how much attention each image's own tokens absorb: for same-sign endpoints, at most two distinct levels allow prompts to follow the shortest path, whereas three or more force a strictly positive extra cost unless the endpoints differ only by rescaling. This gap persists with more prompts, small content perturbations, and aligned multihead outputs. On ViT-B/16 with VTAB-1k, a quadratic Bézier prompt path that minimizes feature action reduces held-out transport action relative to linear and cross-entropy-trained paths between the same endpoints and improves intermediate and online accuracy under unlabeled distribution shift, while cross-entropy-trained paths retain higher minimum accuracy when labeled calibration data are available. Reachable endpoints do not guarantee a reachable shortest path: how prompts move matters, not only where they end. Project page: this https URL.
- [198] arXiv:2611.17826 [pdf, html, other]
-
Title: Space-based Zero-trust Architecture for Securing the Direct-to-Cell ParadigmSubjects: Networking and Internet Architecture (cs.NI)
The rapid evolution of Direct-to-Cell (D2C) technology has emerged as a pivotal enabler for ubiquitous 6G global coverage, seamlessly connecting massive numbers of User Equipment (UEs) in remote and underserved areas. However, by bypassing traditional terrestrial satellite access gateways, the D2C paradigm inadvertently renders conventional front-end security perimeters obsolete. Consequently, malicious UEs can directly expose satellites and their core networks to severe security vulnerabilities. To address these critical threats, this article proposes a Space-based Zero-Trust Architecture (SZTA) empowered by on-board satellite computing. Specifically, the proposed SZTA consists of three core components: radio frequency (RF) fingerprint-based authentication, distributed identity management, and behavior-driven dynamic access control. This architecture enables continuous, real-time verification of UE identity authenticity and behavioral legitimacy, effectively mitigating the inherent network-layer security risks of D2C networks. Simulation results demonstrate the robustness, effectiveness, and efficiency of the proposed SZTA. Finally, to facilitate the practical deployment of {the SZTA}, we discuss key technical challenges, propose corresponding solutions, and outline open research issues to guide future investigations.
- [199] arXiv:2611.17827 [pdf, html, other]
-
Title: SIMTurn: Lightweight Divergent Subgroup Scheduling for Intra-Warp SynchronizationComments: 13 pages, 12 figures, 2 tablesSubjects: Hardware Architecture (cs.AR); Distributed, Parallel, and Cluster Computing (cs.DC); Programming Languages (cs.PL)
Stack-based reconvergence enables efficient SIMT execution with a shared warp program counter and compact control-flow state. However, its fixed serialization of divergent paths can deadlock programs that rely on intra-warp synchronization: a spinning subgroup may prevent a resource-holding sibling from executing. Independent Thread Scheduling (ITS) enables finer-grained progress through per-thread execution state, but introduces additional hardware cost. We present SIMTurn, a lightweight compiler-hardware mechanism that enables forward progress among divergent subgroups while retaining stack-based reconvergence. SIMTurn combines a compiler-inserted yield instruction at potentially blocking loops with a bounded pool of resumable subgroups. The scheduler allows a waiting subgroup to yield to a sibling that can make progress, without requiring independent program counters for every thread. We implement SIMTurn in the compiler, simulator, and RTL on the open-source Vortex GPGPU and evaluate it against baseline IPDOM and per-thread ITS schemes. Across the evaluated GPU thread configurations, SIMTurn introduces only 17-24% scheduler-area overhead relative to IPDOM, compared with 39-91% for ITS, and stays within 8% of IPDOM on workloads that never need cross-subgroup progress. On kernels with genuine cross-subgroup parallelism, SIMTurn tracks ITS closely and is up to 2.7x faster than a software-serialized IPDOM baseline.
- [200] arXiv:2611.17828 [pdf, html, other]
-
Title: Signal Scaffolding: Unlocking the Potential of Lightweight Models for Self-Supervised Monocular Depth EstimationSubjects: Computer Vision and Pattern Recognition (cs.CV)
With the rising demand for edge devices, lightweight self-supervised monocular depth estimation methods have continued to evolve; however, existing methods generally suffer from poor accuracy. We find that the root cause may lies in the failure of conventional training regimes to fully exploit the inherent potential of lightweight models. To address this, we propose a novel three-phase training strategy for lightweight self-supervised monocular depth estimation, termed Signal Scaffolding (SS). Without introducing any additional dataset or parameters, we apply SS to ReForgeDepth, a lightweight network of our own design. Experiments on the KITTI and Cityscapes datasets demonstrate that our method achieves state-of-the-art performance in the accuracy-efficiency trade-off and running in real time on edge devices. These results suggest that lightweight models can be improved further within their existing capacity, and that the central question may be less about how parameters are allocated than about how to unlock this latent potential through training.
- [201] arXiv:2611.17830 [pdf, other]
-
Title: Alleviating Interview Stress: Effects of Robot Persona and Reactive Backchanneling on High-Pressure Mock InterviewsComments: 4 pages, 3 figures, 2 tables. Accepted as a Late Breaking Report (LBR) paper at the 2026 IEEE International Conference on Robot and Human Interactive Communication (RO-MAN)Subjects: Human-Computer Interaction (cs.HC)
High-pressure job interviews induce anxiety and impair performance. Existing mock training suffers from a trade-off among cost, standardization, and physical presence. This study investigated how a robot interviewer's verbal persona (supportive vs. challenging) and nonverbal backchanneling (active vs. negative) affect users' psychological states and facial expressions in a simulated interview. In a 2x2 between-subjects experiment (N=28), participants interacted with a NAO robot. Subjective measures (state anxiety, self-efficacy, social presence) and objective facial emotion recognition were collected. Results showed that a supportive persona reduced task load, increased perceived robot warmth and mind perception, and decreased perceived surveillance and robot discomfort. Active backchanneling increased non-neutral (p = .021) and anxiety-tension facial affect (p = .048). A significant interaction effect between persona and backchanneling was found on social presence (p = .045). The subjective-objective consistency was 67.9%, indicating complementary measurement modalities. The findings provide preliminary support for the social buffering hypothesis in human-robot interaction.
- [202] arXiv:2611.17831 [pdf, html, other]
-
Title: V-JEPA-LM: Video-Language Models That Read Dynamics Beyond Static, Text-Aligned RepresentationsSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
World-model video encoders such as V-JEPA learn how scenes change (direction, speed, order, state) from video alone. To answer questions, they are joined to a language model trained on captions and question-answer pairs. But each caption or question names only a few attributes of a clip, so the model learns to retrieve those attributes rather than to read the clip, and the dynamics the encoder carries go unused: on the same items, a linear probe reads speed from frozen V-JEPA features at 87% while the VLM answers at chance. We show where dynamics lives in a frozen encoder: in a subspace shared across videos for predictive encoders such as V-JEPA, but not as a shared component of frame-wise, text-aligned encoders such as SigLIP. We then build counterfactual supervision that makes the language model read it: pixel transformations turn public videos into answer-balanced variants whose answers are set by the transformation and differ only in the change. Trained in a single stage on 4.53M public samples with 32 visual tokens per frame, where open video LLMs use several stages, 9M to 172M samples and 64 to 256 tokens per frame on average, the resulting V-JEPA-LM models (1B to 8B, Qwen3 and Llama) lead every size class in average score across eleven video benchmarks; our 4B model scores 72.1, against 61.1 for the best open 4B and 64.8 for the best open 8B model. A controlled encoder x supervision study shows that the encoder decides what can be read and the supervision what is read: at a matched budget, counterfactual supervision adds 4.5 points on average and 38 on the speed questions of natural benchmarks. Code and models: this https URL.
- [203] arXiv:2611.17834 [pdf, html, other]
-
Title: PaperCritic: Training AI Reviewers to Assess the Soundness of PapersSubjects: Computation and Language (cs.CL)
As AI conference submissions grow rapidly, AI reviewers are being considered as an alternative. However, simply imitating human reviewing is problematic because reviewing serves multiple purposes, including evaluating papers' soundness, interestingness, and clarity, the last two of which are subjective. In this work, we focus solely on evaluating soundness, the most important and objective criterion, and define the learning objective as jointly maximizing (1) precision, based on the soundness and substantiveness of review points, and (2) recall, the coverage of high-quality reference review points. Under this problem formulation, we construct a quality-labeled review point dataset, PaperSound-23K, using 1K ICLR submissions and 23K human-written and LLM-generated review points. Using this dataset, we train an AI reviewer model, PaperCritic, through reinforcement learning with the F1 score of precision and recall as the reward. Our experiments show that PaperCritic outperforms frontier models and models trained with other AI reviewer training methods on an in-domain test set. Surprisingly, learning to review ICLR submissions also transfers to reviewing Nature-family papers and even to out-of-domain reviewing tasks (e.g., code review, legal contract review, and feedback on story writing).
- [204] arXiv:2611.17836 [pdf, html, other]
-
Title: BLiMP-RO: A Benchmark of Linguistic Minimal Pairs for RomanianSubjects: Computation and Language (cs.CL)
Evaluating grammatical knowledge in language models requires benchmarks that cover diverse linguistic phenomena and are validated against human judgments. We introduce BLiMP-RO, a Romanian minimal-pair benchmark covering 50 grammatical phenomena across ten domains. It comprises 493 linguist-constructed sentence pairs, evaluated by native speakers in five experiments, and 5,942 additional semi-automatically generated, human-validated pairs. We further investigate grammatical generalization under limited linguistic input by training GPT-2 and GPT-BERT models from scratch on Romanian BabyBabelLM data and a newly collected corpus of 10 million words from child-targeted video captions. Increasing caption training data yields approximately 9 percentage points of accuracy improvement per tenfold increase in corpus size. At matched data sizes, BabyBabelLM generally outperforms video captions, while GPT-BERT consistently outperforms GPT-2. Our findings highlight how training data quantity, composition, and model architecture shape grammatical generalization in Romanian.
- [205] arXiv:2611.17845 [pdf, html, other]
-
Title: Tight Quantum Query Bounds for Minimum Cut and Multicut on Dense GraphsSubjects: Data Structures and Algorithms (cs.DS); Computational Complexity (cs.CC)
We determine the dense worst-case quantum query complexity of exact minimum multicut on simple unweighted undirected graphs. With at most $k$ explicit demand pairs, our adjacency-array algorithm uses $\OO(n+\min\{m,n^{3/2}\sqrt{k}\})$ queries; a matching lower bound holds for every $1\le k\le\lfloor n/2\rfloor$ and sufficiently large $n$, with constants uniform in $k$. In particular, the $k=1$ case gives an $\OO(n^{3/2})$ query algorithm for prescribed $s$--$t$ cut. The upper bound combines degree-sensitive spectral estimates of local improvement with refinement by unions of label classes. These steps retain every original optimum and permit exact acquisition without an additional label-count factor. Separately, we prove an $\Omega(n^{3/2})$ global minimum-cut lower bound on dense connected graphs, matching the existing upper bound up to polylogarithmic factors. It holds for the value and for any optimum shore, even with known degrees and sorted neighbor arrays; regular hard instances have balanced optimum cuts. Both lower bounds use degree-preserving switches and classify optimum outputs across a hidden threshold. Our upper bound concerns input queries, not total running time: computation on acquired data may be exponential, even in the single-demand case. Matching bounds at each prescribed sparse edge count are not established.
- [206] arXiv:2611.17846 [pdf, html, other]
-
Title: Is Abliteration Getting Harder? Not Uniformly: Refusal Removal by Direction-Based Ablation and ARA in Heretic on the Qwen3.8, Gemma 4, Olmo 3.1 and Mistral Small 3.2 LLMs Versus Their PredecessorsComments: 76 pages, 19 figures, 34 tables. Code, configurations and judge labels: this https URL ; models: this https URLSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Abliteration, a fine-tuning-free weight edit, removes refusal from open-weight large language models (LLMs). The developer of Arbitrary-Rank Ablation (ARA), an optimisation-based variant, reports that direction-based abliteration, which ablates a refusal direction, increasingly fails on newer models. On five release pairs, refusal removal did not become uniformly harder. Qwen2.5-32B and Qwen3.8-27B, Gemma-2-27B and Gemma-4-31B, OLMo-2-32B and Olmo-3.1-32B, Mistral-Small-2409 and Mistral-Small-3.2, and Qwen2.5-7B and Qwen3.5-4B were edited in Heretic 2.0 by three or four direction-based methods (difference-in-means, MPOA, SOM and, on the small pair, J-space) and ARA under the same prompts, nominal search budget, first-token KL divergence cap and selection rule; removal counted as harder when the automatic WildGuard judge found more of 972 held-out harmful prompts (JailbreakBench, XSTest, StrongREJECT, HarmBench and an author-compiled malicious-code bank) still refused. After direction-based removal, newer releases retained -1 to +37 percentage points more refusal, and after ARA -26 to +34. With run-to-run variance included, between-pair heterogeneity of the direction-based effect ranged from high to low with the scale and variance model (I^2 20-80%). ARA lost less effectiveness than direction-based methods in Gemma and OLMo and more in Qwen; each verdict rests on one ARA run, and some changed with the divergence budget or variance model. In Gemma the direction-based increase came from the malicious-code bank and reversed on public sets. Contrary to prediction, keyword counts at the tool's window did not undercount judged refusal on average, but differed by a mean 7 points per model. Audits of a single pair, method or prompt source can reach opposite conclusions. Whether a newer model resisted removal depended on the release pair, method and budget, not on generation alone.
- [207] arXiv:2611.17850 [pdf, html, other]
-
Title: OrbitWeaver: LLM Agent Planning with Simulation Feedback for Optically Interconnected Orbital GPU ClustersSubjects: Distributed, Parallel, and Cluster Computing (cs.DC); Systems and Control (eess.SY)
Space-based data-center designs increasingly consider GPU computing formations connected by inter-satellite optical links. With limited neighboring connections, jobs assigned disjoint GPUs can still contend on shared relay paths. Rank placement and communication priorities jointly affect this contention and the readiness of downstream computation. We present OrbitWeaver, an offline large language model (LLM) agent planner that jointly optimizes rank placement and communication order. Guided by communication structure, execution dependencies, and simulation feedback, the agent constructs and refines plans that coordinate shared-link access with computation progress. This coordination reduces contention and dependency blocking, shortening mean job completion time (JCT). Simulation experiments on optically interconnected orbital GPU clusters show that OrbitWeaver reduces aggregate mean JCT by 6.2-33.5% relative to the state-of-the-art baselines CASSINI, Crux, and Symphony. These results demonstrate the value of LLM-guided planning for efficient concurrent computing in orbital GPU clusters.
- [208] arXiv:2611.17851 [pdf, html, other]
-
Title: Electrocardiographic Imaging via Posterior Sampling under Operator DiscrepancyComments: Accepted at the GenAI for Health Workshop @ NeurIPS 2026. Code, models and data: this https URLSubjects: Machine Learning (cs.LG)
Electrocardiographic imaging (ECGI) reconstructs heart-surface potentials from body-surface potentials, an ill-posed inverse problem. Diffusion posterior samplers combine a learned prior with an explicit forward operator that they assume to be accurate. In ECGI this operator is only an estimate, because the body changes with respiration and posture, and body composition such as subcutaneous fat is rarely known for an individual patient. The resulting operator discrepancy is structured and depends on the heart signal. We introduce OpDis-PS, a diffusion posterior sampler that models this discrepancy instead of treating it as signal-independent noise. OpDis-PS learns how the discrepancy depends on the signal, folds this into a corrected forward operator, and reconstructs with a fixed prior across all operator conditions. We study two realistic sources of discrepancy, changing lung conductivity and an added subcutaneous-fat layer. Across both, OpDis-PS reconstructs more accurately than classical regularization and learned predictors, with its largest gains on held-out cardiac pathologies.
- [209] arXiv:2611.17852 [pdf, other]
-
Title: PADSNet: Physics-Aware Dual-Stream Network with Biomechanical Priors for Laryngeal Lesion ClassificationZhenzhen You, Weifeng Yu, Zhenghao Shi, Shuangli Du, Minghua Zhao, Yingcheng Zhao, Xinhong Hei, Demin Kong, Xiaoyong Ren, Jin HouSubjects: Computer Vision and Pattern Recognition (cs.CV)
Precise classification of laryngeal lesions is clinically crucial yet highly challenging, as inter-class visual differences are often subtle. While abnormalities in vocal fold vibratory dynamics provide complementary discriminative information, existing deep learning models predominantly rely on visual appearance and lack explicit mechanisms for modeling clinically meaningful vibratory dynamics. To address this, we propose PADSNet, a Physics-Aware Dual-Stream Network for laryngeal lesion classification that integrates visual representations with explicit biomechanical reasoning. Specifically, a Wavelet-Attention Enhancement Module (WAEM) is designed to bridge the spectral-semantic gap, enabling complementary learning of spatial texture and dynamic frequency representations. Meanwhile, a Knowledge Computing Module (KCM) extracts interpretable biomechanical indices from temporal glottal distance waveforms, including indicators of mobility, symmetry, glottal closure, and vibration stability. These descriptors are subsequently incorporated into a Physics-Aware Gating Module (PAGM) as biomechanical priors to adaptively recalibrate visual feature toward pathology-relevant vibratory patterns. Extensive experiments on a clinical dataset demonstrate that PADSNet achieves an accuracy of 0.8452, an F1-score of 0.8090, a sensitivity of 0.7953, and a specificity of 0.8866. These results indicate that explicitly integrating biomechanical priors with deep visual representations provides a more accurate, robust, and clinically interpretable framework for automated laryngeal lesion classification.
- [210] arXiv:2611.17853 [pdf, html, other]
-
Title: DoubleTake: Focused Verification of Visual Evidence for Long-Document UnderstandingSubjects: Computer Vision and Pattern Recognition (cs.CV)
In long-document question answering, a vision-language model can answer incorrectly even when the retrieved pages contain sufficient evidence. Rereading those pages with the original question gives the model little guidance about which fact to check. We introduce DoubleTake, a training-free framework for focused verification of visual evidence. An initial reading produces a candidate answer and a separate evidence record that guides retrieval for a second reading. A planner compares the two candidates and their supporting evidence to formulate a factual question. A verifier answers it from the source images without seeing either candidate or evidence record, and a reasoner uses this observation to resolve the original question. With a frozen Qwen3-VL-8B-Instruct backbone, DoubleTake achieves the highest overall accuracy among the evaluated methods on each of four benchmarks. With the same pages and number of model calls, asking a focused factual question yields higher accuracy on all four benchmarks than asking the original question again. Our code is available at this https URL.
- [211] arXiv:2611.17856 [pdf, html, other]
-
Title: DECODE: Diagnosing Interpretability Failures in Automated Interpretation ProcessSubjects: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
Tools for interpreting the internal representations of Large Language Models (LLMs) increasingly rely on LLMs themselves to both generate and evaluate explanations, creating a self-referential automated interpretation process with minimal human oversight. This automation raises significant reliability concerns, which are exacerbated by the absence of ground truth. In practice, it is difficult to validate explanations or to attribute interpretability failures to their source -- whether they originate from the tool's extracted representations or from the LLM used in the process. To address these challenges, we present DECODE, a visual analytics system that reintroduces human oversight into the automated interpretation process. Our approach decomposes the diagnostic workflow into three progressive stages, isolating failure sources ranging from structural defects to explanation-level misalignments. Drawing inspiration from inter-coder consensus in qualitative analysis, we utilize consensus among independently generated explanations and their evaluation metrics as a proxy for reliability. To further scale human inspection, we employ a three-phase active seeding protocol that strategically recommends high-priority instances for manual review. Finally, we demonstrate the effectiveness of our approach through a case study on 16,384 Sparse Autoencoder features, showing that it can efficiently identify and attribute interpretability failures at scale.
- [212] arXiv:2611.17857 [pdf, html, other]
-
Title: Fuse-All MoE Inference Kernel with Load Balancing on TPUsSubjects: Machine Learning (cs.LG)
Mixture-of-Experts (MoE) architectures activate only a subset of experts for each token, allowing large language models (LLMs) to scale parameter capacity without a proportional increase in computation per token. However, intermediate data movement and uneven expert loads limit the efficiency of MoE inference on TPUs. To address these bottlenecks, we introduce Fuse-All MoE, a TPU kernel design that jointly optimizes expert computation, output communication, and load balancing. The kernel fuses token collection, expert projections, gated activation, and output writes, retaining expert computation intermediates on chip. Fixed token-slot destinations allow scheduling and expert placement to change while preserving routing semantics. In prefill-decode (PD) disaggregated serving, Fuse-All MoE balances EP16 prefill workloads through scheduling and expert redistribution, while symmetric EP8TP2 shares active expert computation between the two TensorCore devices within each chip during decode. Our experiments show that Fuse-All MoE reduces kernel runtime for a measured skewed EP16 prefill workload from 2.3x to 1.09x the same fixed reference time for a balanced workload. Across 32-512 decode tokens, symmetric EP8TP2 reduces kernel latency by up to 21.3% relative to fused EP16. With accompanying serving stack optimizations, the integrated system improves serving throughput by up to 34.5% over the Megablox EP16 baseline.
- [213] arXiv:2611.17858 [pdf, other]
-
Title: AI-aided Decision-Making in Education - Uncertainty, Explainability, and Human ResponsibilityComments: Fabian Kieser and Paul Tschisgale contributed equally and share first authorship. This is a preprint of the following chapter: Kieser, F., Tschisgale, P., Wulff, P., "AI-Aided Decision-Making in Education: Uncertainty, Explainability, and Human Responsibility", published in "AI Roles and Responsibilities in Education", edited by S. Papadakis, 2025, Springer, reproduced with permission of SpringerSubjects: Artificial Intelligence (cs.AI)
As Artificial Intelligence (AI)-based systems have become increasingly integrated into various aspects of everyday life, including education, understanding how these systems make decisions is of paramount importance. AI holds significant potential to enhance educational practices through the facilitation of personalized and adaptive learning, automation of assessment procedures, and assistance with administrative tasks. However, the opacity of AI-based decision-making in education (especially with deep learning and generative AI models) can hamper decision-making, especially when outcomes appear unjustified or incorrect. In this chapter, we argue that three key principles (uncertainty, explainability, and human responsibility) are essential for the legitimate use of AI in educational settings. Acknowledging the inherent uncertainties in AI-based decisions, analogous to those in human decision-making, is fundamental. Explainability is critical for ensuring that the functioning and outcomes of AI applications are transparent and comprehensible to both educators and learners. Furthermore, human responsibility is crucial, particularly in human decision-making, which remains indispensable across all stages of AI development and deployment. By examining these three principles, this chapter provides a comprehensive guide for stakeholders seeking to integrate AI thoughtfully and effectively into the educational system.
- [214] arXiv:2611.17859 [pdf, html, other]
-
Title: Per-Agent-Safe Navigation of Globally Driven Robot Collectives via Affine Field DesignComments: 9 figures, 1 tableSubjects: Robotics (cs.RO)
Safe navigation of robot collectives requires coordinating motion and deformation so that the collective can traverse constrained environments while ensuring each robot's safety. Existing controllers for swarmalator collectives largely rely on local control, where each robot receives an individual command, which causes the control and communication requirements to grow with the number of agents. We introduce a broadcast control formulation in which a single spatially affine field governs both the motion and deformation of the collective. The scalar component of the field regulates collective size, and its spatial gradient generates collective translation through non-reciprocal interactions. Both components enter the collective dynamics directly, which allows per-robot wall, actuation, and cohesion constraints to be enforced using control barrier functions within a single quadratic program for the whole collective. The resulting controller continuously adapts the collective size to the local environment, and enables safe passage through corridors narrower than the unconstrained collective without pre-planned resizing. We further derive how the available actuation budget determines the reachable speed and minimum passable corridor width, and we demonstrate that two independently driven collectives can simultaneously traverse a shared junction using distinct fixed phase labels without inter-collective coordination.
- [215] arXiv:2611.17862 [pdf, html, other]
-
Title: FlexCast: Flexible Context Learning for Time Series Foundation ModelsZuyang Shen, Luqi Gong, Fanda Fan, Bingsong Wang, Xiang Ao, Yahui Wei, Minghao Liu, Chunjie Luo, Jianfeng ZhanComments: 15 pages, 8 figures. Code: this https URLSubjects: Artificial Intelligence (cs.AI)
Time series foundation models (TSFMs) based on autoregressive (AR) next-token prediction entangle two factors: variable-context conditioning and prediction-unit factorization with autoregressive rollout. Strong AR performance therefore does not reveal whether generalization stems from the flexible training distribution, the local prediction target, the left-to-right inference process, or their combination. We propose FlexCast, a flexible context-target learning framework that disentangles these factors. FlexCast reformulates forecasting as variable-length masked prediction over time series patches: during pretraining, it dynamically samples visible context and masked target patches, exposing the model to diverse context-target configurations without sequential decoding. At inference time, the same pretrained model supports one-step forecasting, confidence-based iterative refinement, and next-token-style generation through a unified mask-and-predict interface, enabling controlled comparison of inference schedules under identical architecture, data, and objective. Evaluated on BLAST and multiple zero-shot benchmarks, FlexCast achieves accuracy competitive with or better than strong TSFM baselines while substantially reducing inference latency. Ablations confirm that context diversity and contiguous target masking drive performance, while left-to-right decoding does not yield consistent gains once flexible context-target relationships have been learned, suggesting that the key benefit of AR-style TSFMs lies in flexible training rather than sequential inference. The code is available at this https URL.
- [216] arXiv:2611.17863 [pdf, html, other]
-
Title: Implicit discretization of PID-like sliding-mode control with third-order output accuracyComments: Extended version of a paper submitted to Automatica. 20 pages, 3 figures, 2 tables. Code and verification records: this https URLSubjects: Systems and Control (eess.SY)
A digital PID-like sliding-mode controller is proposed for a double integrator with a Lipschitz matched disturbance. A proper implicit discretization, matched to the exact sampled dynamics of the plant, attains the accuracy of the continuous-time controller without discretization chattering. Global finite-step convergence to a discrete-time sliding mode is established for two gain designs and every sampling time, using an implicit Lyapunov function and a computer-assisted certificate in exact arithmetic. The resulting controller provides third-order output accuracy, second-order accuracy for its derivative, and exact finite-time rejection of constant matched disturbances. These accuracy orders are proved to be optimal among deterministic causal controllers using the same sampled information.
- [217] arXiv:2611.17864 [pdf, html, other]
-
Title: Disentangling Forgetting from Utility: Lightweight LLM Unlearning via Orthogonal Null-Space ProjectionXiuyuan Wang, Weiming Liu, Zongxin Yang, Fan Wang, Liemiao Jin, Rubin Qin, Wei Chen, Bin Shi, Chaochao Chen, Xiaolin ZhengComments: 12 pages, 4 figuresSubjects: Machine Learning (cs.LG)
Large Language Models (LLMs) pose significant security risks by retaining private or hazardous information, creating an urgent need for unlearning techniques that excise specific knowledge without compromising general utility. However, prevailing fine-tuning-based unlearning methods struggle to achieve this balance, often suffering from computational inefficiency and unintended utility degradation due to the entangled nature of knowledge within model parameters. To disentangle forgetting from utility, we propose UUE, a theoretically grounded framework that reformulates unlearning as controllable model editing. First, UUE projects parameter updates into the orthogonal left null space of retained knowledge representations, strictly ensuring that the editing process causes minimal interference to the model's general capabilities. Second, distinct from targeted mapping, we introduce a convex untargeted objective that derives a closed-form analytical solution to effectively disrupt undesired knowledge by pushing forget-set embeddings toward their antipodal points. We implement UUE via lightweight, pluggable adapters (including a low-rank variant, UUE-L), enabling efficient deployment without modifying the frozen backbone. Extensive experiments on TOFU and WMDP benchmarks demonstrate that UUE and UUE-L achieve superior unlearning efficacy, significantly outperforming existing methods.
- [218] arXiv:2611.17865 [pdf, html, other]
-
Title: Starlight: Whittle-Driven Fidelity Rescue for Satellite-Terrestrial Quantum NetworksComments: 10 pages, 7 figures, 2 tablesSubjects: Networking and Internet Architecture (cs.NI)
A planetary-scale quantum internet is constrained by exponential attenuation in terrestrial fibers. Low-Earth-orbit (LEO) satellites provide a low-loss vacuum backbone, but limited satellite resources must be allocated among competing connections under continuously evolving and heterogeneous physical conditions. Existing satellite scheduling remains largely provisioning-oriented, optimizing satellite-resource assignment for aggregate entanglement delivery. Recent age-based approaches instead incorporate chronological freshness into entanglement scheduling. Neither line of work directly identifies the connections most urgently requiring satellite assistance as their physical states evolve. We propose Starlight, a demand-driven satellite-terrestrial entanglement routing architecture that targets degrading connections before their terrestrial entanglement becomes unusable. Starlight introduces Effective Age of State (EAoS), a physics-aware metric that maps heterogeneous decoherence and swapping-induced fidelity penalties onto a unified additive coordinate. Representing each connection by evolving EAoS, we formulate rescue scheduling as a Restless Multi-Armed Bandit (RMAB) problem and derive a closed-form Whittle-style index approximation for top-K selection. Starlight further employs a two-timescale control plane and an asynchronous dispatch-and-mask pipeline to suppress redundant in-flight allocations. Simulations show that Starlight achieves up to 84.5% average satisfaction under constrained rescue capacity and sub-0.5-ms orchestration at 1,000 connections, while remaining robust to heterogeneous memory decay and stochastic atmospheric degradation.
- [219] arXiv:2611.17866 [pdf, html, other]
-
Title: Active Learning of Temporal Specifications through Distinguishing QueriesSubjects: Logic in Computer Science (cs.LO); Artificial Intelligence (cs.AI)
Temporal specifications are essential for verification and monitoring, yet writing formulas that capture the intended behaviour is difficult. Active learning derives such specifications through interaction between a learner and a teacher, but remains underexplored for temporal logics. Existing approaches require the teacher to answer equivalence queries and to construct counterexamples, and they often rely on SAT-based learning procedures that restart the search after each new example. We present an active-learning framework for LTLf in which the teacher only labels traces and answers natural-language questions about the intended behaviour. The learner maintains an enumerated version space of candidate formulas, generates traces that distinguish candidates, and uses the teacher's hints to rank candidates or to restrict the search. We prove consistency with the observed examples, identification within bounded hypothesis classes with a bound on the number of queries, and a PAC guarantee for the final hypothesis. In experiments with a simulated teacher on 36 targets, including targets in four standard environments, the learner identifies every target when it compares all rivals, with fewer than seven queries beyond an initial sample of ten traces on average. Correct hints reduce the number of queries by up to 16% and the number of generated candidates by a fifth, and hints used for ranking never prevent identification, even when incorrect. Our enumeration-based version space is faster than the SAT-based approach of previous learners at every target size.
- [220] arXiv:2611.17867 [pdf, html, other]
-
Title: UltraBERT-JEV: Encoder models for System One ApplicationsSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
We study encoder models that pick one of several options in a single forward pass, without generating text. This technical report compares recipes for fine-tuning UltraBERT-380M with one mask per option. The four recipes differ in how an option is scored and how the scores become probabilities; one follows the expert gate of DeepSeek-V3. Trained on the same data as Qwen3-0.6B with the LLM-as-Jev recipe (67.6% test accuracy), the encoder reaches 69.7% to 72.1%. Scoring an option by "yes" minus "no" instead of "yes" alone gives the larger gain, from 69.7% to 71.3% on the test set and from 57.8% to 59.6% on external benchmarks (56.1% for the decoder). Sigmoid normalisation adds 0.8 test points; a square-root softplus adds none. One request at a time on an NVIDIA A100, the encoder takes 29 ms per decision against 68 ms, but uses more peak memory (2.8 against 2.2 GiB). An auxiliary head, trained on human-marked evidence and on labels selected with a reward adapted from SelfCite, predicts in the same pass the sentences annotated as support for the answer.
- [221] arXiv:2611.17868 [pdf, html, other]
-
Title: Speed-Variation Distillation for Generative ModelingComments: CDEL-ECCVW 2026 proceedings. Project page: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV)
While diffusion models excel at image generation, their denoising trajectories pass through corrupted intermediate states far from the image manifold. We ask whether generation can instead proceed through recognizable images throughout the entire trajectory. We introduce a generative framework built around a variation-reducing degradation process that contracts diverse images toward a shared prototype while preserving recognizable intermediate states. Because such manifold traversal lacks a simple analytical form, we construct the trajectories using generative-prior-based image morphing. We then distill the reverse process into a diversification operator that transforms the prototype into increasingly varied samples. To resolve the one-to-many ambiguity of the reverse process, we associate each trajectory with its initial noise latent and apply perturbations to improve robustness against accumulated errors. At inference, every step yields a recognizable image, allowing early termination and enabling a controllable speed--variation trade-off.
- [222] arXiv:2611.17869 [pdf, html, other]
-
Title: Process Mining-Enabled Federated Anomaly Detection for the Industrial IoTComments: Accepted to the IEEE Annual Congress on Artificial Intelligence of Things (IEEE AIoT 2026)Subjects: Machine Learning (cs.LG)
While federated anomaly detection enables the integration of local data insights across diverse nodes to reduce false positives, it inherently struggles with non-IID data. In practice, industrial IoT plants often operate under varying conditions and experience diverse environmental phenomena, resulting in significant heterogeneity in the monitored time series. In such settings, standard federated learning approaches fail to perform nuanced aggregation that accounts for these cross-node differences, ultimately leading to increased false positive rates. In this paper, we propose a process mining approach to evaluate the similarity of local node data against a reference provided by a global node. The global node utilizes these similarity metrics to dynamically weight local models during aggregation, producing personalized updates distributed across the network. We evaluate our approach using industrial IoT time-series data collected from manufacturing plants operating under distinct conditions. Experimental results demonstrate that our process mining-enabled federated anomaly detection pipeline achieves superior performance, yielding a 46.52% and 11.99% relative improvement in F1-score compared to isolated local training baselines and standard FedAvg, respectively, alongside a 51.24% relative reduction in false positives over standard federated aggregation.
- [223] arXiv:2611.17870 [pdf, html, other]
-
Title: OmniScore-Bench: Evaluating Vision-Language Models on Musical Score ReadingSubjects: Artificial Intelligence (cs.AI)
Vision-language models (VLMs) can answer questions about musical scores, but their ability to recognise notation, recover its structure and interpret its musical meaning requires separate evaluation. Existing optical music recognition (OMR) resources establish detection and transcription measurements; a transcription score alone leaves the strengths and bottlenecks of a VLM score reader unresolved. We introduce OmniScore-Bench, an evaluation of OMR capabilities in VLMs through 3,000 questions spanning six aspects and 24 task subtypes. Each question specifies a visible scope, reference annotations and an answer schema, enabling automatic evaluation from symbol localisation to musical consistency and evidence-linked question answering. Our assessment of twelve VLMs reveals operation-specific capability profiles and different transcription leaders across excerpt, single-page and multipage tasks. Controlled analyses examine the roles of score images, auxiliary givens and additional page context. On the tested subset, image removal exposes visual dependence in pitch and repeat interpretation while measuring how much performance remains with public context alone. Separate format and content analyses distinguish answer-format violations from musical errors. OmniScore-Bench supports diagnosing which score-reading operations VLMs perform reliably and which require improvement.
- [224] arXiv:2611.17872 [pdf, html, other]
-
Title: Protection-Oriented Assessment of Machine Learning-Based Fault Detection in Power SystemsJulian Oelhaf, Georg Kordowich, Changhun Kim, Christian Bergler, Andreas Maier, Johann Jäger, Siming BayerComments: 9 pages, 3 figures, 5 tables. Accepted for oral presentation at the 21st International Conference on Developments in Power System Protection (DPSP APAC 2027), Hong Kong, 6-9 January 2027Subjects: Machine Learning (cs.LG); Signal Processing (eess.SP)
Machine learning (ML)-based fault detectors are increasingly proposed for power system protection, yet aggregate metrics can obscure the consequences of false and missed trips. We present a protection-oriented assessment of three ML detectors and a sequence-current reference on PROTECT-90, comprising 9,022 simulated fault episodes on a 90 kV double-circuit line. The assessment combines five window conditions, causal window-end labelling, episode-grouped cross-validation, and operating-point selection under a fixed security budget. It reports security, dependability, detection timing, decision margins, and episode-level nuisance assertions. Gradient-boosted trees, a convolutional neural network, and logistic regression achieve AUC-ROC values above 0.999, yet at a security budget of $10^{-3}$ their developed-fault missed-trip rates range from 0.007% to 1.99%. Post-clearing false-trip rates of approximately 0.12-0.16% per window coexist with nuisance assertions in 11.3-19.1% of episodes. Among the analysed physical descriptors, missed trips are most consistently associated with higher fault resistance within the sampled range. An analytical current proxy suggests a fault-current gradient only partly reproduced by waveform-derived current (rank correlation 0.62). No single topology state or faulted phase consistently dominates within the sampled domain. Because PROTECT-90 contains fault episodes only, security is assessed on pre-fault and post-clearing windows rather than independent non-fault disturbances.
- [225] arXiv:2611.17874 [pdf, html, other]
-
Title: What a Memory Preview Cannot Tell You: Ambiguity in AI Memory PermissionsComments: 4 pages, 2 figures, 1 table. Code and data: this https URLSubjects: Human-Computer Interaction (cs.HC)
A memory rule states when an assistant may reuse a remembered event, and an editor can confirm a rule by previewing one use it allows and one it blocks. We ask what such a preview lets a person infer about the rule they saved. For an editor with six or twelve valid rules, we enumerate every rule compatible with a preview. Splitting the rules most evenly does not target the question a user cares about. For an explicit preference, two adaptive tests leave all three recommendation scopes open (for a one-off choice the outcome turns on a tie), while one test of transfer to another task settles the question. Weak previews are also the default. Under the intended rule, 36 of 39 possible allowed/blocked pairs contain a blocked example that no rule allows, and no pair identifies the rule. Three or four tests suffice here; when permissions depend on each other, a fixed checklist grows to six or nine and adaptive questioning needs at most five or seven. These counts use a specified rule vocabulary and fictional events. In a small pilot, 10 of 20 people who saw such a weak preview of a rule they had just chosen mispredicted, or were unsure about, at least one of two uses, and none of those 10 said the outcomes matched what they wanted. Whether better previews help people is untested.
- [226] arXiv:2611.17876 [pdf, html, other]
-
Title: Delayed Effects of Temporary Weight Decay in Muon PretrainingMd Rifat Arefin, Ravid Shwartz-Ziv, Ernie Chang, Chinnadhurai Sankar, Rylan Conway, Aristide Baratin, Irina Rish, Adithya Sagar, Patrick HuberSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
A temporary optimization intervention can increase validation loss when it ends yet reduce it at the final training budget. We demonstrate this with temporary weight decay in Muon-trained decoder-only Transformers with 162M, 405M, 1.4B, and 3.4B parameters. For models up to 1.4B parameters, moving a fixed-duration decay window from early to late training reverses its final effect. At 162M, changing window duration alters final loss even at approximately matched total shrinkage. Most strikingly, a schedule selected at 162M and transferred without retuning to 1.4B is 0.26 nats worse than no decay when removed, overtakes it only after 73% of training, and ultimately finishes 0.028 nats better while also outperforming the best sampled constant-decay baseline. Early decay leaves a persistent weight-norm deficit, which keeps relative steps larger after the intervention ends. An approximate norm-dynamics recurrence captures the slow convergence of these relative steps, and fixed-norm replays of the measured trajectories reproduce the performance ordering and a similar loss difference. At 162M and 405M, norm-reset experiments show that smaller weight norms at the end of the decay window explain most of the final loss improvement. A separate Gram intervention exhibits related post-removal norm and relative-step dynamics. These results show that temporary optimization schedules can have delayed effects long after removal and should therefore be evaluated at the intended final training budget, not only at removal.
- [227] arXiv:2611.17878 [pdf, html, other]
-
Title: DySENSE: Dynamic Spatiotemporal ENergy Synthesis and Emulation with Satellite ImageryComments: Accpted by KDD 2027Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Cities account for over two-thirds of the world's energy and more than 70% of global carbon dioxide emissions. City energy is essential for achieving the 2050 net-zero target and the UN's Sustainable Development Goals 7 and 11. Although existing studies have made progress, they are limited by data and methodological gaps. Existing studies are largely predictive for existing cities rather than generative for future cities; fine-grained, aligned monthly dynamic energy data are scarce; and existing GenAI models struggle to generate dynamic urban functional monthly energy. We present DySENSE, a joint multimodal GenAI framework that synthesizes static urban satellite imagery together with spatially aligned, dynamic monthly building energy consumption maps. We decouple long-term static urban morphology from short-term dynamic energy consumption. By conditioning on static urban land use, DySENSE can generate realistic, diverse and new static urban imagery with aligned energy maps. By conditioning on dynamic monthly and climate prompts, DySENSE can jointly generate dynamic, monthly energy consumption maps while keeping the urban morphology static. Experiments across four cities (New York, Washington, Norfolk, Busan) demonstrated that DySENSE can generate annotated high-fidelity synthetic energy data using less than 25% labelled data, boosting downstream prediction performance by 8% mIoU, and achieving strong energy dynamic consistency. Compared to existing methods, DySENSE significantly improved prediction performance (8.7% accuracy) and reduced prediction error (9%-18% CVRMSE, ASHRAE). DySENSE offers an effective foundation for energy-efficient cities, peak-load analysis, building performance simulation and dynamic demand response for city, energy, and building science. The data and code: this https URL and this https URL
- [228] arXiv:2611.17879 [pdf, html, other]
-
Title: From Local Prediction Error to Final Outputs: A Theory of Expert Sharing in Multi-Step ReasoningComments: 96 pages, 3 figures, 28 tablesSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Mixture-of-experts (MoE) models can share experts across layers or computational steps, pooling training information from different contexts. This can reduce local prediction error, but multi-step systems are ultimately evaluated by their final outputs. We study when these local gains survive composition. In specified linear regression models, we derive exact sharing and merging criteria by balancing variance savings against bias from sharing across roles with different target behavior. Under a common fixed linear propagation map, we characterize expected squared final-output risk, retaining error directions and correlations. We construct cases where sharing lowers average local error yet increases final-output error. Within this setting, we also identify when orthogonal full pooling with isotropic noise preserves every local improvement at the final output. For general routed computations, we derive conditional end-to-end error bounds that account for distribution shifts, errors in surrounding components, and changes in expert selection. Separately, we show that correct predictions and route labels alone do not establish that individual experts implement the attributed operations. Controlled experiments realize a strict local-to-final risk reversal. In low-data symbolic settings, trained Transformers exhibit statistically supported teacher-forced one-step gains without statistically resolved free-running final-output gains. Separate intervention tests distinguish successful multi-layer actions from single-expert attribution and generalization to unseen transitions. Together, these results distinguish when sharing improves local estimation, when those gains survive to the final output, and what evidence is needed to attribute reusable operations to individual experts.
- [229] arXiv:2611.17880 [pdf, html, other]
-
Title: Soteria: Formally Verified Planning with Runtime Enforcement for Safe LLM AgentsComments: Accepted as NeurIPS'26 OralSubjects: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
LLM agents that execute multi-step tool calls must satisfy two objectives simultaneously: completing the user's task (utility) and conforming to domain policies and correctness constraints (safety). These objectives are in tension -- blocking an unsafe action prevents a violation but can leave the agent stranded with no principled way to recover. Existing approaches sacrifice one objective for the other. We introduce Soteria, a framework that reconciles safety and utility through verified hierarchical planning and runtime enforcement of specifications. Before execution, the agent generates a structured plan that is formally verified against the specifications prior to any tool invocation. During execution, the verified plan serves dual roles: it ensures that the agent takes specification-conformant trajectories and provides guidance when unsafe actions are blocked. Across multiple benchmarks covering a diverse range of tool-use tasks, Soteria achieves 1.7 times the average task success rate of the best baselines, with no violations of policy specifications, demonstrating that safety and utility need not be traded off.
- [230] arXiv:2611.17881 [pdf, html, other]
-
Title: Photogrammetric 3D Model-Based Animal Behavior ClassificationOlaf Hellwich, Moritz Heine, Patrik Reiske, Sole Traverso, Katharina Hohlbaum, Henning Sprekeler, Christa Thöne-Reineke, Lars Lewejohann, Roman BhuiyanComments: 8 pages, 9 figures, 1 table, conference proceedingsJournal-ref: GFaI Future Tech Days Tagungsband 2026, pp. 56 - 63, Gesellschaft zur F\"orderung angewandter Informatik e.v., Berlin 2026, ISBN 978-3-942709-38-5Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Video-based behavior estimation is often implemented end-to-end from image sequences. We advocate a two-stage alternative: reconstruct 3D body and object geometry before classifying behavior. This makes ambiguities of the ill-posed 2D-to-3D inverse problem explicit and potentially correctable. We demonstrate the strategy with simulated interacting mice and extend it to real video.
- [231] arXiv:2611.17883 [pdf, html, other]
-
Title: Securing Multi-Agent Systems by DesignSubjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Multi-agent systems solve complex tasks by decomposing them into subtasks, delegating the subtasks to agents, and aggregating the agents' responses using an aggregation rule. Malicious agents can manipulate the final solution by providing malicious responses to their assigned subtasks. In this paper, we first propose response manipulation attacks (ReMA), which iteratively optimize malicious responses to maximize attack effectiveness. Specifically, we formulate untargeted and targeted attacks as optimization problems and develop an approach to approximately solve these problems by iteratively refining malicious responses based on feedback from a surrogate multi-agent system. To defend against such attacks, we further propose SecMAS, a secure aggregation rule that leverages the semantic similarities among agent responses to select the response that is most consistent with a sufficiently large subset of the other responses. We theoretically establish security guarantees for SecMAS against both untargeted and targeted attacks. Extensive evaluations demonstrate that existing aggregation rules are vulnerable to ReMA, whereas SecMAS effectively defends against all evaluated attacks, including adaptive variants of ReMA, while maintaining utility comparable to that of existing aggregation rules.
- [232] arXiv:2611.17886 [pdf, html, other]
-
Title: Quantum Graph Neural Networks: Architectures, Trainability, and ApplicationsSubjects: Machine Learning (cs.LG)
Quantum graph neural networks (QGNNs) apply parametrized quantum circuits to graph-structured data, with the aim of overcoming limits of classical graph neural networks in scalability, expressivity, and long-range information flow. Since the first Hamiltonian-based model in 2019, the field has produced many architectures, applications, and claims, but the evidence behind them has not been assessed systematically. This survey introduces a unified formalism that describes every QGNN as an encoder, a graph-conditioned circuit, observables, and an optional classical head, together with a circuit-level condition for permutation equivariance and a resource model for comparing costs. We organize the literature into five architecture families (convolutional, equivariant, message-passing and recurrent, Hamiltonian, and hybrid) and compare them by qubit cost, symmetry, scalability, and evidence. We then review symmetry and geometric inductive biases (permutation, rotation, and Lorentz), the results on expressivity, trainability, noise, and scalability, and applications in high-energy physics, chemistry, finance, healthcare, earth observation, and transportation. An audit of datasets and evaluation practices leads to a minimum reporting protocol. Our main finding is that the design space is rich and symmetry is the best-supported inductive bias, whereas the evidence remains thin: results are mostly simulated, on small graphs, at parity with classical models or with fewer parameters, and no advantage has yet been established on tasks over classical graph data. We identify five open challenges and a research agenda covering provable advantage, equivariant and geometric designs, fault-tolerant implementations, and shared benchmarks. The coverage extends through 2026 and emphasizes recent work.
- [233] arXiv:2611.17887 [pdf, other]
-
Title: RuralShield: Super Intelligence-Assisted Security Triage and Serverless Infrastructure for Resource-Constrained HealthcareComments: 13 pages, 2 figures, 6 tables. Code: this https URLSubjects: Cryptography and Security (cs.CR)
Resource-constrained healthcare providers must connect cloud security controls to an operational response process. Introducing language-model assessments adds an authorization problem: a recommendation must be converted into a permitted action against a specific resource. RuralShield combines a configurable serverless application foundation with structured model assessment, deterministic routing, and human-approved containment. Its design separates workload access, assessment validation, and action authorization so that each boundary produces inspectable records. We evaluate the implementation through a frozen 120-hour non-AI workload and separate desktop and AWS AI cohorts. The workload recorded 999,861 authenticated requests and 999,860 successful responses, with one HTTP 404 and one missing minute representing 139 scheduled requests. Median, 95th-percentile, and 99th-percentile Ohio client latencies were 28.57, 53.01, and 110.77 ms. Eight successful model requests exercised assessment and reporting in two four-case cohorts. The AWS cohort and five scripted controls produced nine completed workflows, including two human-approved Lambda containment cases with verified restoration of the original concurrency setting. Source-linked, checksum-verified archives preserved the records before scoped cleanup. The study demonstrates an inspectable integration of application controls and human-mediated AI response under the evaluated synthetic conditions.
- [234] arXiv:2611.17889 [pdf, html, other]
-
Title: PLL-Less Synchronization of Grid-Following Converters Using a Voltage-Driven Frenet FrameSubjects: Systems and Control (eess.SY)
In grid-following converters, the synchronization path can transmit voltage disturbances to the current loop under weak or unbalanced grid conditions. This paper develops a Frenet-frame controller that derives its axes from the integral curve of the measured voltage at the point of common coupling. The tangential current governs active power, while the normal and binormal currents govern the transverse components of non-active power. Feedforward terms derived from the output filter model compensate curvature and torsion coupling, yielding identical nominal current dynamics across the axes. The controller operates without a phase-locked loop or sequence extraction. A reduced transverse model accounts for the finite bandwidth of geometric rate reconstruction and predicts a reduction in low-frequency coupling from voltage disturbances to current. Controller hardware-in-the-loop tests using real-time simulation of a three-phase four-wire converter show the shortest settling time after an active power reference step and the smallest peak power deviations during balanced voltage disturbances among the tested controllers. Under single-phase voltage unbalance, the method reduces peak binormal current and maximum phase-current total harmonic distortion relative to both benchmarks. It also settles to the commanded power in a weak-grid connection test where the phase-locked loop benchmark exhibits sustained oscillations.
- [235] arXiv:2611.17892 [pdf, html, other]
-
Title: Hierarchical Bayesian physics-augmented neural networks for populations of hyperelastic laws: specimen variability and uncertainty from scarce, noisy dataComments: 41 pages, 17 figures, 7 tablesSubjects: Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Machine Learning (stat.ML)
The hyperelastic law of a material such as soft tissue differs between specimens, and often only a few specimens can be tested. Physics-augmented neural networks (PANNs) learn such laws from data without a prescribed functional form while keeping them physically admissible, and Bayesian versions of related models quantify the uncertainty of the law of one specimen or of a pooled group of specimens. A pooled fit, however, attributes the differences between specimens to measurement noise or to uncertainty about a mean law, and neither describes a new specimen. We first develop a Bayesian PANN for one specimen, whose priors keep the learned law plausible for few, noisy data. We then extend it to a hierarchical Bayesian PANN, a nonlinear mixed-effects model in which all specimens share learned polyconvex features and each specimen is a draw from a population of laws. Features, specimens, and population are inferred jointly by variational inference, with a separate model of the measurement noise, so that the variability between specimens is not mistaken for noise, the population carries its own uncertainty, and every specimen drawn from it is polyconvex. The population predicts the response of new specimens with credible intervals, without data or from a partial test, and assigns them to cohorts with quantified uncertainty, all without a prescribed form of the law. On synthetic populations, it recovers the population from few specimens and predicts new specimens more reliably than three published methods. On Treloar's rubber and on healthy and aneurysmal aortas, it reaches the accuracy of a deterministic PANN and assigns no aorta specimen to the wrong cohort with confidence, although it underestimates the uncertainty of untested curves.
- [236] arXiv:2611.17897 [pdf, html, other]
-
Title: See and Detect: Image-Driven Open-Vocabulary Multi-View 3D Object DetectionComments: tech reportSubjects: Computer Vision and Pattern Recognition (cs.CV)
Recent open-vocabulary multi-view indoor 3D detection methods typically follow a point-cloud-driven pipeline, where pseudo 3D boxes are generated from reconstructed or estimated geometry and semantic labels are inferred from projected regions. However, such a design makes box generation dependent on the completeness of point clouds and restricts semantic learning to local appearance, causing inferior 3D object localization and open-vocabulary classification. In this paper, we present See and Detect, an image-driven framework that uses multi-view images as the primary evidence for 3D box and semantic label generation. For box generation, we first discover object candidates from individual images, enabling objects missing in point clouds to be recovered from visual evidence. However, single-view predictions are often noisy and redundant. To address this, we propose multi-view consensus box consolidation to fuse spatially consistent predictions across views, and box quality assessment to reduce the impact of boxes with partial object coverage. For semantic label generation, we introduce context-aware semantic learning, where semantic labels are queried from full image context, enabling more accurate semantic prediction. Extensive experiments on ScanNet and ARKitScenes demonstrate that our method achieves state-of-the-art open-vocabulary multi-view 3D detection performance while running over 500$\times$ faster than the previous state-of-the-art method.
- [237] arXiv:2611.17898 [pdf, html, other]
-
Title: Evaluating Polling Schemes for OFDMA Uplink Scheduling in Time-Sensitive Wi-Fi NetworksSubjects: Networking and Internet Architecture (cs.NI)
IEEE 802.11ax introduced orthogonal frequency division multiple access (OFDMA) to improve channel utilization through concurrent multi-user transmissions. For effective uplink (UL) scheduling, the access point (AP) requires timely knowledge of buffered traffic at the stations (STAs). Explicit buffer-status polling provides this, yet each polling round can query only as many STAs as there are resource units (RUs). Scheduled-access, random-access, activity-based, and learning-based schemes address this buffer-status acquisition problem differently and have generally been evaluated in isolation under different conditions, so it remains unclear how they compare across time-sensitive workloads. We evaluate common schemes from the Wi-Fi standard and literature under three time-sensitive environments: a large-scale teleconferencing session, an industrial IoT deployment, and an extended reality (XR) service. The relative performance of the schemes depends strongly on traffic characteristics and offered load. In conferencing, activity-based and learning-based polling perform best, while conventional contention-based access remains competitive under light load. In the industrial scenario, scheduled-access most effectively delivers time-sensitive packets within their deadlines under heavy load, whereas contention-based access performs best under lighter load. In XR, OFDMA-based schemes achieve high video throughput that contention-based access cannot sustain; among them, activity-based polling achieves the lowest deadline-miss ratios for uplink head-tracking traffic, at the expense of longer video delay tails. These findings highlight the importance of selecting channel access and scheduling schemes according to traffic requirements and operating conditions.
- [238] arXiv:2611.17899 [pdf, html, other]
-
Title: More Genes, Faster Inference: Efficient Large-Panel Spatial Transcriptomics Prediction from HistologyComments: Accepted at KDD 2027Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Quantitative Methods (q-bio.QM)
Spatial transcriptomics (ST) prediction from histology aims to infer spatially resolved gene expression profiles directly from tissue images. This is a challenging problem that is currently often addressed with complicated model architectures, additional inputs, and complicated training procedures, while remaining limited to fixed, small gene panels. This raises a natural question: Is there an efficient method for large-panel ST prediction? Here, we present GeneFlow, a simpler yet effective framework for large-scale spatial gene expression prediction that unifies flow-matching generative modeling with structured multimodal fusion. Specifically, GeneFlow first encodes both histological morphology and cross-gene interactions in spatial spots through histology and gene tokenizers. The model directly predicts spatial gene expression, rather than an intermediate velocity field, while being trained with a velocity-based flow-matching objective. We show that, even without relying on heavy generative pipelines, GeneFlow achieves state-of-the-art performance across multiple ST benchmarks, with up to 56.8% gains in 1,000-gene prediction. Extensive experiments further show that GeneFlow produces biologically meaningful spot profiles while reducing computational cost by orders of magnitude, thereby improving efficiency and scalability, particularly in clinical settings. We believe GeneFlow provides an opportunity for leveraging widely available histology images for downstream biological discovery. Code is available at: this https URL
- [239] arXiv:2611.17900 [pdf, html, other]
-
Title: AmbiSteer: Selectively Eliciting Ambiguity-Aware Responses in Vision-Language ModelsComments: 23 pages, 23 figuresSubjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
Vision-language models (VLMs) often exhibit overconfidence, providing firm answers even when visual evidence is insufficient. A prominent case is referential ambiguity in Visual Question Answering (VQA), where multiple entities in an image are plausible referents of a question. Addressing such cases requires recognizing when ambiguity is present: ambiguous questions can be handled by asking for clarification or enumerating plausible referents, while unambiguous questions should receive a direct answer. Current VLMs struggle with this distinction, often defaulting to a single interpretation. To investigate whether activation steering can serve as an inference-time mechanism to elicit such behaviors, we introduce AmbiSteer, a lightweight framework to selectively elicit ambiguity-aware behavior. Across three VLMs, we find that steering alone has limited and model-dependent effects, while prompting is generally more effective. Combining prompting and steering substantially increases clarification and enumeration behaviors, but also elicits them on unambiguous cases. AmbiSteer addresses this using a linear classifier to predict input ambiguity and condition the steering direction accordingly, increasing the proportion of inputs handled appropriately from 59.0% to 77.0% on average across models. Our work sets the ground for lightweight approaches to reduce model overconfidence in uncertain multimodal settings.
- [240] arXiv:2611.17903 [pdf, other]
-
Title: Zoom In, Scale Out: Metric Depth Models Mistake Digital Zoom for Moving CloserComments: 22 pages including supplementary material. Code: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV)
Many new AI models estimate how far away things are, in metres, from a single photo. To do this they must also work out what camera took the photo. We check whether they really do with a simple test: digital zoom, that is, cropping a photo and enlarging the crop back to full size. Zooming moves nothing in the scene, so a correct model should give the same distance for every point still in the picture. The test is free, exact and needs no labels. Two numbers describe a model's reaction: its depth changes like $s^{-\alpha}$ (ideal $\alpha=0$) and its focal-length estimate like $s^{\beta}$ (ideal $\beta=1$), for zoom factor $s$. On 800 images from KITTI, NYUv2 and DIODE, the widely used Depth Anything V2 has $\alpha=0.95$: it treats a zoomed photo almost as if the camera had moved closer, because it judges distance mainly from how large things look. Models that also predict the camera remove only part of this error ($\alpha\approx0.31$-$0.36$), because their focal estimate grows too little. Given the true camera, Depth Pro and DA3-Nested keep their depth ($\alpha\approx0$), so the error lies in their camera estimate. The reason: a zoomed photo is, pixel for pixel, a normal photo of a scene squashed in depth, so size alone cannot tell the two apart, and a focal estimator trained by regression stays close to the cameras it saw in training. Undoing this needs no labels or training and makes Depth Pro pass the test ($\alpha=0.04$) at almost no cost on normal photos. Controls, including a real optical zoom lens, rule out blur and resizing artefacts. Finally, label-free fine-tuning for self-consistency lowers $\alpha$ to 0.05 on KITTI and 0.07 on NYUv2 once a "blurry means zoomed" shortcut is removed; we report what this costs on normal and unseen photos.
- [241] arXiv:2611.17905 [pdf, html, other]
-
Title: Cracks in Embeddings: Robustness and Stability of JEPA-Based Chest X-Ray RepresentationsSubjects: Computer Vision and Pattern Recognition (cs.CV)
Self-supervised Joint Embedding Predictive Architecture (JEPA) models are increasingly used for medical imaging, however there is a limited understanding about robustness of these models toward real-world image changes. We test four JEPA-based chest X-ray encoders, trained on 78,468 NIH ChestX-ray14 images, under six tests: two black-box adversarial attacks (SPSA, Boundary Attack), three image corruptions (Gaussian noise, brightness shift, contrast shift), and a label-scarcity test, measuring embedding change and classification AUROC. These perturbations cause severe, model-dependent damage: encoder-level Cosine Similarity drops as low as 0.33, and AUROC falls by as much as 0.12 in the worst case (I-JEPA under the Boundary Test). We test two ways of training the small predictor on top of each frozen encoder to address this. Adversarial fine-tuning, trained on clean and perturbed images together, raises embedding consistency above 0.97 in nearly every case and improves AUROC, relative to the same perturbed baseline, for most models, with the largest gain of $+0.10$ on I-JEPA. A Vector-Quantized codebook method, trained only on clean images and evaluated on perturbed test images for the first time, instead degrades under distortion for all four models, most noticeably for I-JEPA ($-0.15$ AUROC on average), showing that its discrete bottleneck alone does not provide the robustness that comes from training on perturbed data directly.
- [242] arXiv:2611.17907 [pdf, html, other]
-
Title: Game-Theoretic Analysis of RIS-Assisted Challenge-Response PLA - Part II: Partial Channel KnowledgeComments: Part II of arXiv:2611.16557. Submitted to IEEE TIFSSubjects: Cryptography and Security (cs.CR)
We consider a physical layer authentication scenario in which a legitimate user (Alice) communicates with an authenticator (Bob) through a reconfigurable intelligent surface (RIS), controlled by Bob. An attacker (Trudy) aims to impersonate Alice by suitably precoding her messages and still transmitting them through the RIS. This two-part study analyzes this attack-defense interaction between Bob and Trudy under perfect (Part I) and partial (Part II) channel knowledge by all the agents. Part I formulates the joint RIS selection and precoder design under perfect channel knowledge as a finite zero-sum game, establishing exact channel subspace conditions for attack indistinguishability alongside equilibrium mixed strategies. Part II extends the defense to incomplete channel information, where each agent observes only its own channel, by formulating a finite-action Bayesian game and characterizing the resulting Bayesian Nash equilibrium. Simulations in geometric and 3GPP TR 38.901 UMi-NLOS channels reveal that authentication performance is governed by available spatial degrees of freedom, channel correlation, and operating signal-to-noise ratio. The proposed equilibrium strategies effectively balance communication rate and security, outperforming conventional benchmark policies.
- [243] arXiv:2611.17908 [pdf, html, other]
-
Title: DiffusionSSD: Achieving Extreme Visual Compression with Dual-stage Diffusion and Cross-level Near-Data ProcessingComments: 15 pages, 19 figuresSubjects: Hardware Architecture (cs.AR); Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
Growing visual data overwhelms storage on resource-constrained edge devices, making compression essential. Traditional codecs offer inadequate rate-distortion performance, motivating the exploration of diffusion codecs. Diffusion codecs face two limitations: (1) At the algorithm level, diffusion decoders preserve perceptual quality under extreme compression, but require substantial computation. (2) At the architecture level, in NPU-centered devices, neural codecs occupy the NPU, interfering with other AI applications. Meanwhile, the complex codec pipeline incurs long I/O paths, leading to additional overhead and energy cost.
To address these issues, we propose DiffusionSSD, a cross-level near-data processing framework for extreme visual data compression on edge devices, using an algorithm-architecture co-design method. Specifically, (1) at algorithm level, we present Dual-stage Diffusion, a lightweight solution tailored for diffusion codecs that employs patchification and single-step diffusion, reducing computation cost while preserving perceptual quality; (2) at architecture level, we introduce a Cross-level Near Data Processing architecture that integrates the proposed computational universal flash storage (cUFS) and LPDDR-PIM, enabling codecs to be entirely offloaded to data side while reducing visual-data movement and avoiding SoC redesign; (3) at system level, combining both the algorithm and the underlying architecture, we propose a Patch-Workload co-aware scheduling scheme to enable more efficient system-level execution. Experiments show that Dual-stage Diffusion maintains comparable perceptual quality while reducing computation cost by up to 92.7%. Compared with NPU+DRAM and near-data-processing design such as SmartSSD, cUFS+PIM improves coding throughput by 2.32 times and 2.91 times on average, respectively. - [244] arXiv:2611.17909 [pdf, html, other]
-
Title: CardioTrace: Motion-Aware Cine Synthesis via Coupled Tracing and SegmentationSubjects: Computer Vision and Pattern Recognition (cs.CV)
Cardiac cine MRI provides time-resolved visualization of cardiac motion and is essential for evaluating ventricular function and anatomy. In routine clinical acquisition, however, cine sequences are frequently temporally sparse or degraded by arrhythmia, motion, and scan-time constraints, which can hinder reliable quantification and downstream analysis. We present CardioTrace, a motion-aware cine synthesis framework that generates a full cardiac cine sequence from limited input while explicitly modeling and tracing cardiac motion. CardioTrace couples two complementary pathways: a pixel-translation branch that directly predicts cine frames in a residual formulation, and a flow-field branch that estimates time-resolved deformation fields from a reference phase to provide interpretable motion correspondences across frames. To further preserve anatomical plausibility, a segmentation branch supervises clinically meaningful cardiac structures at key phases, and a cross-branch consistency objective encourages the synthesized appearance dynamics to follow the explicit motion guidance, mitigating temporal drift and unrealistic shape transitions over long sequences. The framework also supports prompt-style conditioning to incorporate auxiliary patient or acquisition information, improving robustness across heterogeneous multi-center data. Experiments on public cardiac cine MRI datasets demonstrate that CardioTrace improves perceptual quality and temporal coherence over representative GAN- and diffusion-based baselines, while better preserving ventricular morphology throughout the cardiac cycle. \textit{Our codes are available after review}.
- [245] arXiv:2611.17910 [pdf, html, other]
-
Title: Adversarially Attacking Symbolic Vocabulary Vulnerabilities In LLM PlannersComments: Published at NeurIPS 2026Subjects: Artificial Intelligence (cs.AI)
As large language models (LLMs) are increasingly deployed as autonomous agents, their abilities in navigating automated planning tasks have become essential. However, despite their strengths, LLMs' planning ability is not robust, particularly when they are not allowed to exploit commonsense cues in the symbolic vocabulary (e.g., names of actions, predicates, and objects) used to define planning tasks. This makes LLMs unreliable compared to symbolic planners. With recent reasoning models showing improved planning performance on tasks with obfuscated symbolic vocabularies, we utilize an adversarial attack framework to reveal to what extent the crucial weaknesses of LLM agents at joint symbolic interpretation and planning remain and whether they can reason properly when presented with severe cases. As such, our main contribution is devising an LLM-based attack model, which we call Symbolic Swapper, to automatically generate obfuscated planning domains with non-standard symbolic vocabularies that systematically break strong reasoning models' ability to plan. In doing so, we reveal how much LLMs' planning abilities can be further degraded under more advanced obfuscation schemes, allowing us to ascertain their success when having to rely on their pure planning ability. We find that attacks can decrease the planning performance of base models and even robust planning frameworks by over 70%. We further analyze the factors behind how LLM planning abilities break down and find that successful obfuscation causes models to appear to significantly degrade in reasoning quality despite them showing some inherent ability to understand the domain, revealing a gap in the perception of a task domain and being able to plan under it.
- [246] arXiv:2611.17914 [pdf, html, other]
-
Title: DOCKET: A Pluralistic Benchmark for Evaluating AI-Assisted Policy Feedback in U.S. RulemakingAlexander Spangher, Charlotte Peale, Sílvia Casacuberta, Victor Y. Wu, Mirac Suzgun, Olivia H. Martin, Dan Bateyko, Cassandra Handan-Nader, Diyi Yang, Sanmi Koyejo, Daniel E. HoSubjects: Computers and Society (cs.CY)
As AI models advance and impact deliberative democracies, there is a growing need to assess their support of pluralistic viewpoints. Most pluralism benchmarks are insufficient for this: they are not grounded in diversity (i.e., stakeholders across different frames, skills and needs), quality (i.e., rigorous and technical debates), and impact (i.e., invested deliberators and real-world outcomes). To address this, we first (i) curate DOCKET, the most comprehensive dataset of the U.S. Federal Notice & Comment process to date. DOCKET consists of 42.3M public comments on 22,400 regulations across 44 government agencies between 2016 and 2026 and contains millions of policy debates; we link public comments to particular policy items proposed by regulators to track the public's attention and the responsiveness of regulators to public input. (ii) Next, we demonstrate the pluralistic utility of DOCKET - we show that public commenters have a wide breadth of perspectives with significant depth in policy expertise and that regulators are often responsive. (iii) Finally, we present DOCKETBench, the first systematic evaluation of LLMs' ability to represent human policy perspectives in federal rulemaking. Only 18-30% of LLM-generated feedback aligns with real human comments. LLMs critique proposals broadly but generally, lacking the specific grounding, citations, and lived-experience arguments that humans bring to bear; in addition, most models we test systematically agree with government perspectives (overlapping with up to 68% of their claims) and agree least with labor unions (as little as 26% of their claims), demonstrating a fundamental skew in perspectives that even tools like persona modeling do not solve. Taken together, we reveal fundamental gaps in how language models navigate real policy issues, and DOCKETBench serves as a testbed for closing those gaps in a crucial real-world domain.
- [247] arXiv:2611.17915 [pdf, html, other]
-
Title: SWIFT: Sample-Wise Instance Fine-Tuning of Time-Series Foundation Models for ForecastingSubjects: Artificial Intelligence (cs.AI)
While current few-shot adaptation strategies for time series forecasting foundation models are predominantly dataset-centric, the necessity of obtaining an additional dataset for the finetuning process poses a significant hurdle for adoption. To address this constraint, we propose Sample-Wise Instance Fine-Tuning (SWIFT), a novel forecasting adaptation framework that achieves fine-grained specialization by leveraging structural priors from the target instance's own history. This shift allows the fine-tuning process to be entirely transparent to the end-user, which significantly enhances the ease of adoption. To implement this framework, SWIFT synergizes Self-Retrieval for historical mining and Frequency-Aware Loss for spectral consistency, while employing Deviation-Aware Dynamic Fusion to balance adaptation with general knowledge preservation. Extensive experiments across four TSFM backbones and three long-history benchmark subsets validate SWIFT, yielding average MASE and WQL reductions of 6.41% and 4.83% across 24 combinations of four backbones and six Chronos Benchmark II datasets. To the best of our knowledge, SWIFT is the first framework to investigate sample-wise instance fine-tuning of time-series foundation models, specializing the model for an individual prediction window using only the corresponding forecasting sample's own observed history.
- [248] arXiv:2611.17917 [pdf, html, other]
-
Title: When Cooperation Must Change: Semantic Strategy Unlearning in Multi-Agent Reinforcement LearningSubjects: Artificial Intelligence (cs.AI)
Cooperative multi-agent reinforcement learning (MARL) supports coordinated decision-making in distributed Web services, multi-robot control, and collaborative games through learned coordination conventions. When deployment requirements change, a previously useful convention may become obsolete or unsafe. For example, revised cache-sharing permissions in Web content delivery may require caches to abandon a learned division of storage and serving roles while maintaining efficient delivery. Teams must therefore remove the specified convention while preserving useful cooperation. This is challenging because the same relational pattern can be reproduced by different agents across different histories. Runtime shielding leaves the underlying coordination preference intact and can impair task progress, while existing reinforcement unlearning and trajectory-deletion methods do not directly target role-invariant conventions. This paper formalizes this problem as semantic coordination-strategy unlearning and proposes this http URL uses causal coordination attribution to identify influential interactions and retain-aware subspace updates to limit interference with unrelated behaviors. Counterfactual re-coordination helps agents learn productive alternatives, while adversarial auditing tests whether the target convention can still be recovered. Accepted updates are merged into decentralized actors for direct deployment. Evaluation across five cooperative MARL benchmarks shows that CARVE reduces mean nominal convention-completion and worst-case elicitation rates by 91.4\% and 83.3\%, this http URL median utility loss is 5.9\%, and measured inference latency matches that of the original policies.
- [249] arXiv:2611.17918 [pdf, html, other]
-
Title: Condensing Law of Security InformationSubjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Scaling laws indicate that the generalization error of a large language model (LLM) decreases monotonically with model size once data is sufficient, and the corollary that larger models outperform smaller ones is now widely accepted. We argue that this corollary fails in security: decisions there hinge on anomalous inputs, which are long-tailed (e.g., zero-day attacks) and therefore poorly served by likelihood fitting. Worse still, the capability brought about by scale (e.g., inferring the intended whole from a corrupted part) leads a LLM silently \emph{repair} the anomaly before any downstream security head ever sees it. We formalize this as a tradeoff between two quantities: the cognitive capability $K(W)=I(W;P)$ that a model holds about the world distribution, and the security-information retention $R(W)=\frac{I(Z;S)}{I(A;S)}$ that survives in its intermediate representation. We derive three information-theoretic bounds to connect them: an upper bound on restoration accuracy in terms of $K(W)$, an upper bound on $R(W)$ in terms of the restoration error, and a Fano lower bound on security-task error in terms of $R(W)$. These theorems approximately imply that usable security information is a product of how much security information survives and how well it can be extracted, indicating that optimal usable security information can be achieved at a certain model size. We state this as the \emph{condensing law of security information}: usable security information is unimodal in model size, so security generalization error is U-shaped and is minimized at a finite $N^{\star}$ rather than decreasing without bound. The experiments on various real-world LLMs validate the condensing law, indicating that there is an optimal model size to solve security issues.
- [250] arXiv:2611.17919 [pdf, html, other]
-
Title: Vessel2Brain: Edge-Aware Multi-Scale Flow Matching for 3D Brain Image Generation from Vessel MasksSubjects: Computer Vision and Pattern Recognition (cs.CV)
Generating realistic 3D brain MRI volumes from sparse anatomical priors is a highly ill-posed yet clinically valuable problem, with important applications in data augmentation, simulation, and downstream neuroimaging analysis. In this work, we investigate the challenging task of synthesising full 3D brain MRI volumes conditioned solely on vascular label maps. While vascular structures provide strong anatomical and topological cues, they occupy only a small fraction of the brain volume and offer no direct information about surrounding tissue appearance. To address this challenge, we propose Vessel2Brain, an edge-aware multi-scale conditional flow matching framework that explicitly incorporates vascular structure information into the generative process. Our approach formulates vessel-guided brain image synthesis as a conditional continuous flow, enabling stable training and flexible generation while preserving structural correspondence. A multi-scale flow matching strategy is introduced to jointly model global anatomical coherence and fine-grained local details, with auxiliary velocity predictions providing additional supervision across spatial resolutions. Furthermore, we design an edge-aware objective that emphasises vascular boundaries during training, encouraging anatomically faithful tissue synthesis around vessel structures. We evaluate Vessel2Brain on two public 3D brain MRI datasets with paired vascular annotations. Extensive quantitative and qualitative experiments demonstrate that our method consistently outperforms state-of-the-art generative models in terms of reconstruction accuracy, structural consistency, and distributional realism.
- [251] arXiv:2611.17921 [pdf, html, other]
-
Title: RecMem: Recurrent Memory Compression for Long-Sequence Generative RecommendationComments: Accepted at NeurIPS 2026Subjects: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
Generative recommendation benefits from long interaction histories, but the cost of processing them directly grows with history length. We propose RecMem, a recurrent memory compression framework guided by three desiderata: space invariance, semantic fidelity, and lifelong evolvability. RecMem compresses historical segments into a fixed set of memory slots using a bifurcated encoder-decoder with shared lower layers and slot-wise forget-gated updates. Recommendation supervision trains the memory to preserve predictive information, while stream-consistency training aligns predictions from sequential and packed processing. By construction, the memory state and decoder context are independent of history length; under contraction and approximate semigroup consistency, we further derive a horizon-independent bound on the state discrepancy between these two processing orders. On MerRec, RecMem uses 16 memory slots and a 216-token decoder context, compared with 1,200 tokens for HSTU-1200, while retaining 87.9% of HSTU-1200's Recall@10 and 97.0% of its Recall@200. It outperforms PHSTU and the evaluated recent-only and retrieval baselines, and analyses of distant history show that archived interactions contribute predictive signal.
- [252] arXiv:2611.17924 [pdf, html, other]
-
Title: TIDE: Reclaiming Lost Performance In On-Policy Reinforcement LearningMathew Jacob, Morgan Borjigin-Wang, Kan Zhu, Po Hao Chen, Yi Pan, Andy Ruan, Nayani Modugula, Ethan Lavi, Luis Ceze, Malte Schwarzkopf, Akshay Narayan, Deepti Raghavan, Stephanie WangComments: 15 pages, 9 figuresSubjects: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF)
Synchronous, colocated RL post-training of large language models (LLMs) today suffers from severe under-utilization due to long-tail rollouts. Inference engines fall idle waiting for long-tail rollouts or decode at low, inefficient batch sizes. To address this, we present TIDE, an RL training system that preserves on-policy training semantics while recovering resource utilization. TIDE does so by two complementary techniques: (1) dynamically repurposing rollout GPUs for training while long-tail rollouts complete and (2) dynamically migrating rollouts between inference engines to boost decoding throughput. TIDE automatically determines the right time to repurpose rollout resources through an adaptive migration policy, obtaining performance improvements that generalize across multiple applications and hardware configurations without requiring bespoke configuration. In our evaluation, TIDE provides end-to-end speedups of up to 1.36x over colocated systems, outperforming all evaluated on-policy baselines and achieving up to 97% of optimal throughput.
- [253] arXiv:2611.17925 [pdf, other]
-
Title: What Balanced Accuracy Cannot Identify in Neural-Signal Decoding under Test-Time AdaptationSubjects: Artificial Intelligence (cs.AI)
Adaptation for neural-signal classification is usually transductive: each batch is scored after updating on it, and the gain depends on the adapted model [4]. This paper asks what an aggregate change fails to characterise inside the cohort. Full test-time adaptation (full_ttt) was evaluated in 11 dataset-encoder cells (four each on Siena/EEG and PTB-XL/ECG, three on Sleep-EDF/EEG), with splits and source-model evaluation fixed; the subject was the unit, exact sign tests were Holm-corrected and conditional on the shared stream, and bootstrap intervals accompanied a dependence audit. Two results of different kinds follow. In the binary case balanced accuracy gives a one-class decoder the same value as one reporting both classes at 50%, so the aggregate cannot identify a class that stopped being reported; a class covering a tenth of a subject's samples went unpredicted in 26-47 of 112 subject-cells, against 7-9 for the controls, which is a descriptive count. With cohort, split and update rule fixed, the within-subject class contrast follows the encoder, tested directly rather than inferred from tests against zero: two of six Siena encoder pairs survive the dependence readings at their nominal reference and one at a fold-level reference, while two of ten Sleep-EDF stage pairs pass the run-clustered gate at its nominal reference; aggregate and contrast named different encoders in 10 of 11. The protocol bounds both results, since a stream-interleaving effect reverses sign with state sharing (+3.5 to -2.4 pp). An aggregate score can hide individual harm, as [1] showed, but it cannot distinguish a uniform loss from a deleted class; class-level change should therefore be reported alongside the aggregate and subject-level change alongside it, with both results conditional on the cell, the rule and the stream construction.
- [254] arXiv:2611.17929 [pdf, other]
-
Title: Evaluating Large Language Models for Scientific Paper Search: Retrieval Success and Inference Latency on Mac Studio and NVIDIA A100Frederic Andriantseheno Rakoto (Mnemosyne), Yannis Bendi-Ouis (Mnemosyne), Xavier Hinaut (Mnemosyne)Subjects: Artificial Intelligence (cs.AI)
Scientific literature search can benefit from locally deployed LLM assistants, which reduce reliance on hosted model providers while giving research teams greater control over their data and infrastructure. This work investigates whether open-weight models can support web-assisted paper retrieval under practical computing constraints. We first characterize inference latency as the input context grows from 512 to 128k tokens on a Mac Studio and a server equipped with two NVIDIA A100 GPUs. We then evaluate eight models on the retrieval of 70 papers from seven scientific repositories, using queries that represent three levels of domain knowledge. The results show that retrieval success depends strongly on the selected model and the specificity of the query. And that A100 GPUs are generally faster than the Mac Studio, but its unified memory allows larger models that could not be loaded in the tested GPU configuration.
- [255] arXiv:2611.17930 [pdf, html, other]
-
Title: Token2Tile: Token-Wise Sparse Attention with Tile Execution through a Low-Rank Space for Video DiTsComments: 21 pages, 14 figures, 23 tablesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Attention is essential for modeling spatio-temporal dependencies in video diffusion transformers, but its quadratic cost makes it a major inference bottleneck. Its high sparsity offers an opportunity to accelerate generation by focusing computation on important interactions. However, accuracy and execution favor different granularities: each query attends to its own key tokens, whereas GPU kernels work on tiles. We present Token2Tile, a training-free sparse attention operator that combines per-query scoring and a per-query residual with tile execution. An online low-rank space, computed directly from queries and keys to preserve attention logits, makes clustering, scoring, and residual approximation efficient. Token2Tile unites query-specific selections to allocate exact work across query clusters, then packs the selected keys into GPU tiles. On MiniMax-H3 and Wan2.1-14B, Token2Tile delivers 1.48-1.54x end-to-end speedup over dense FlashAttention-3 with VBench scores within 0.5 points of dense. At comparable speeds on H3, it improves PSNR over SOL and SubBlock by 5.5-6.8 dB; on Wan, it exceeds SOL by 4.6-4.8 dB while running faster.
- [256] arXiv:2611.17931 [pdf, html, other]
-
Title: MineralExp: Benchmarking LLM Agents for Experimental Design in Real-World Mineral ExplorationBing Wang, Mingyang Wang, Yuanhang Mi, Yue Xu, Kaibo Su, Li'ang Xiao, Chunguo Wu, Changchun Li, Ximing Li, Kun-Feng Qiu, Bo YangComments: 16 pages, 6 figuresSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE)
In mineral exploration, verifying geological hypotheses relies on designing a variety of geochemical experiments. While large language model (LLM) agents increasingly automate downstream geoscientific data analysis, a crucial upstream capability remains largely unexplored: automated geochemical experimental design, i.e., deciding which experiments to conduct before committing to costly laboratory work. Evaluating this capability of LLM agents is difficult for two reasons. First, geochemical experiments exhibit high methodological diversity and subjectivity: to test the same hypothesis, different geologists may choose different minerals and instruments, so no single protocol can serve as the reference answer. Second, open-ended experimental plans resist representation as structured labels, which makes objective scoring difficult. To address these challenges, we introduce MineralExp, a benchmark that reformulates geochemical experimental design as a hypothesis-based experiment selection problem, i.e., a multi-label classification problem. We abstract concrete instrument names and procedural details into a shared taxonomy of 19 experimental types across five mineral-system processes: Source, Transport, Deposition, Modification, and Preservation. This taxonomy yields deterministic multi-label targets while accommodating diverse experimental protocols. Constructed from published ore-deposit literature and verified by domain experts, MineralExp comprises 277 cases and 1,572 reference experimental plans. Given a geological hypothesis and initial context, an LLM agent selects the relevant tasks and generates concrete experimental plans, which are scored by multi-label classification metrics and an LLM-as-a-Judge metric. The experimental results show that current agents generally achieve high recall but low precision, tending to propose more experiments than are necessary to test the hypothesis.
- [257] arXiv:2611.17933 [pdf, html, other]
-
Title: OutfitterBench: Evaluating LLM Agents on Long-Horizon Resource Allocation and Commitment CoordinationSubjects: Artificial Intelligence (cs.AI)
In many long-horizon settings, an agent's own choices create future constraints: committing finite capacity establishes future obligations and forecloses later opportunities. Sustaining performance therefore requires both keeping evolving plans feasible and comparing the value of resource allocations with the alternatives they displace. Recent benchmarks increasingly evaluate LLM agents in continuing operations, but give less attention to how agents keep their own commitments feasible while reassessing resource allocations as opportunities change. We introduce OutfitterBench, where agents manage an outdoor-equipment rental business for 200 days, purchasing from a 46-product catalog and handling roughly 1,100 reservation requests. Under a fixed capital ceiling, reservations and uncertain walk-in demand compete for reusable equipment whose availability changes over time. Experiments with seven frontier LLMs show that several agents underperform a simple rule-based baseline, while only the leading agent consistently surpasses it. Trajectory analysis shows that some agents apply known constraints inconsistently and make new decisions that conflict with existing commitments. Agents with more reliable fulfillment can still secure low returns from comparable commitments or delay resource reallocation. The leading agent better connects preparation with economic choices but also misses opportunities for more profitable asset use. These findings highlight limitations in sustaining coherent resource plans and reassessing their value over time. The benchmark code is publicly available at this https URL.
- [258] arXiv:2611.17937 [pdf, html, other]
-
Title: Tracing Actual Causes with Counterfactual Witness MapsComments: Accepted at NeurIPS 2026 (Main Track)Subjects: Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
Halpernian actual causation, the formal study of which events in a specific situation caused a specific outcome in acyclic structural causal models, underpins legal responsibility, moral blame, and the assessment of causal harm. Despite the rich logical framework, no prior sound and complete enumeration procedure is known for finding all actual causes in a given setting, blocking downstream tasks such as responsibility attribution, blame assessment, and causal-harm grading that aggregate over the full cause set. We close this gap by introducing witness targets: sets of joint configurations containing every counterfactual witness. In combination with an ancestor intervention grammar we derive a finite directed acyclic witness map. We prove that for every actual cause there exists a contingency set such that the joint intervention set appears as a node of this map, and present an algorithm for Tracing Actual Causes (TrAC) that is sound and complete for enumerating all actual causes of any target event. We compare TrAC empirically with existing single-cause baselines on random Boolean structural causal models, demonstrating practical feasibility for graphs of up to 15 nodes.
- [259] arXiv:2611.17938 [pdf, html, other]
-
Title: Measured from the feed: the unsustainable growth of AI video generationComments: Main paper + Supplementary paper Under submission at a journalSubjects: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Computers and Society (cs.CY); Emerging Technologies (cs.ET)
Text-to-video generation runs at commercial scale, but its broader environmental impacts remain, to our knowledge, understudied. Providers rarely disclose video volumes or energy needs; we measured the yearly video output of two commercial platforms and built a prospective model of such services' absolute footprint from training to storage. Our projections suggest that reducing emissions depends on both limiting growth in video generation and achieving substantial efficiency gains. Under continued growth, emissions rise nearly seventy-five-fold by 2030 despite technological improvements. When usage remains at 2026 levels, the same improvements reduce emissions by about two-fifths. These estimates are a conservative lower bound: they cover one AI service among many; other services and uses would push the total higher.
- [260] arXiv:2611.17939 [pdf, html, other]
-
Title: Can You Trust Your Finance Agent? Drift, Diagnosis and Auditable Intervention on One Shipped Financial Research AgentComments: 9 pages. Accepted as an oral presentation at the AI4F workshop, ICAIF 2026, MilanSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Can you trust a finance agent you did not build? We take the reference agent shipped with the Vals Finance Agent Benchmark v2, leave it untuned, and ask four questions a leaderboard score does not answer. Is the score stable? Seven identical runs on 18 questions pass between 105 and 123 of 149 rubric criteria, and the bill varies by up to 17%. When the agent fails, who owns the fix? We classify 20 failed trajectories, from three runs of nine held-out questions, by the layer whose owner would remedy them, under a taxonomy frozen in advance: 13 belong to the agent's instructions, 3 to its tools, 4 to the model, none to its data. Can that diagnosis be trusted? Four blind readers from three model families, given only the traces, name the same fault as our LLM analyzer in 94% of cases, the rate at which they agree with one another (93%). Can a change be trusted? An optimization loop reads the traces, makes one targeted change per round and keeps it only if it clears a fixed threshold. Under a cost objective its final configuration costs 0.34x the baseline on the held-out questions (95% CI [0.22, 0.58]) at accuracy within noise. Under an accuracy objective it gains 14 points on the development questions and 1.5 on the held-out ones (CI [-5.3, +7.2]). Judged against the measured run-to-run noise, under a rule written before the test was opened, the first is a real saving and the second is not. This is a case study of one agent and nine test questions, not a general result.
- [261] arXiv:2611.17944 [pdf, html, other]
-
Title: Differentially Private Federated Class-Incremental Learning for EEG Mental State ClassificationSubjects: Machine Learning (cs.LG)
Federated learning (FL) under differential privacy (DP) is typically evaluated on large client cohorts, where subsampling and cross-client averaging amplify privacy and smooth injected noise. Personal biosignal applications such as wearable EEG instead involve small federations of a handful of subjects, a regime whose interaction with formal client-level DP is largely uncharacterized. We study differentially private federated class-incremental learning in this small-cohort regime, combining a public EEG dataset with a custom four-channel MUSE 2 recording set into a K=5-client federation, and give a masked per-class training protocol that avoids degenerate gradients and missing negative supervision under class-incremental streaming. We propose FedEEG, combining Synaptic-Intelligence consolidation with a lagged adaptive FedProx term, and show it reduces catastrophic forgetting relative to FedAvg, FedProx, and a federated Learning-without-Forgetting baseline; on-device balanced-batch replay reduces mean forgetting by a further 56.7%. A class-scoping diagnosis yields $50.5\%\pm3.6\%$ accuracy with 6 seeds and 2.02 times chance. We also introduce head-scoped client-level DP, which protects the classification head and alone reduces the achievable $\varepsilon$ by 200 times at a matched round budget; combined with fewer communication rounds it reaches $\varepsilon\!\approx\!10.4$ with utility not statistically distinguishable from non-private training.
- [262] arXiv:2611.17947 [pdf, other]
-
Title: A narrow band method for non-causal discretizations of anisotropic eikonal equations, with adaptive width, sub-and super-solution construction, and quasi-linear complexityJean-Marie Mirebeau (CNRS, CB), Yanis Mouheb (CB, CEA/DAM)Subjects: Numerical Analysis (math.NA)
We describe a narrow band method for numerically solving anisotropic eikonal equations, and more generally static first order Hamilton-Jacobi equations with possibly non-convex Hamiltonians, discretized using standard schemes (semi-Lagrangian, upwind-differences, Lax-Friedrichs). The causality property is not needed, in contrast with the fast marching method. The proposed approach simultaneously builds a numerical sub-solution u and super-solution u to the problem, and adapts the width of the narrow band based on their comparison. We prove that (1 + $\epsilon$)u $\ge$ u at termination, where $\epsilon$ > 0 is any given tolerance parameter, and that the worst case complexity is O(N | ln $\epsilon$| ln(N/$\epsilon$)), where N denotes the number of discretization points, and the underlying constant is explicit in terms of those associated with the scheme properties. Numerical experiments illustrate our results.
- [263] arXiv:2611.17950 [pdf, html, other]
-
Title: Toward Interference-Agnostic RF Source SeparationComments: 5 pages, 3 figures, 1 tabs. Submitted to ICASSP2027Subjects: Artificial Intelligence (cs.AI); Signal Processing (eess.SP)
Separating a radio-frequency signal of interest from a co-channel interference is mostly addressed with specialized separators, each trained for one pair of signal and interference types and therefore relying on a label that a heavily jammed receiver cannot identify. We show that this reliance is not benign: applied to an interference it was not trained on, such a separator usually fails to recover the signal, and in most of the cases we measure it leaves the signal worse than no processing at all. We then train a single separator over every pair, giving it no indication at test time of which one it faces, on two architectures with no component in common: the convolutional WaveNet of the ICASSP 2024 RF Challenge, and a band-split transformer that we adapt from music source separation. Both behave the same way. At the size of one specialized model, the unified separator stays close to the whole models it replaces, keeping the bit error rate close on almost every pair, better on one and worse on another. On an interference absent from every training set, it does as well as the best specialized model, without anyone having to choose which one to deploy.
- [264] arXiv:2611.17951 [pdf, html, other]
-
Title: Parameterized Approximation for Minimizing Uncovered Elements and Unsatisfied ConstraintsSubjects: Data Structures and Algorithms (cs.DS)
Min-UnCoverage asks for at most $k$ sets minimizing uncovered weight. Min-Unsat SAT and Min-Unsat CSP ask for an assignment with at most $k$ true variables minimizing the weight of unsatisfied clauses or constraints.
For constraints of arity at most $d$, we give a deterministic $(1+\varepsilon)$-approximation for weighted Min-Unsat CSP in $(dk/\varepsilon)^{O(k)}N^{O(1)}$ time, where $N$ is the input size. This also gives approximation schemes for bounded-arity SAT and coverage with bounded element frequency. We show that weighted coverage with set size at most $d$ can be solved exactly in $2^{O(kd)}N^{O(1)}$ time by adapting color-coding. For coverage on incidence graphs of bounded semi-ladder index, we give a deterministic parameterized approximation scheme .
We show that, unless $\mathrm{FPT}=\mathrm{W}[1]$, SAT and CSP have no finite-factor FPT approximation parameterized by $k+d$ when variable frequency is at most $d$, or on $K_{3,3}$-free incidence graphs of semi-ladder index at most four. Known Hitting Set hardness gives the same lower bound for all three problems at VC dimension two.
Under the same assumption, we show that no computable relaxation $g(k,d)\geq k$ allows an FPT algorithm to preserve a prescribed unsatisfied-weight target for SAT or CSP, already for arity three and, separately, bounded variable frequency. For coverage, known algorithms preserve the target using at most $k+1$ sets on incidence graphs of bounded semi-ladder index. - [265] arXiv:2611.17952 [pdf, html, other]
-
Title: Domain Adaptation Based Pipeline for Character Classification and Handwritten Text RecognitionJournal-ref: International Journal on Document Analysis and Recognition (IJDAR), 2026Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Character classification and handwritten text recognition, particularly in specialized tasks, often suffer from limited labeled data, while a large amount of unlabeled data remains unused. To address this challenge, we propose a novel unsupervised domain adaptation-based pipeline to reduce the distribution gap between source and target domains. Our pipeline follows three main steps. In the first step, we pretrained an expert model on source data, to learn a strong general representation. In the second step, we perform unsupervised domain adaptation to leverage unlabeled target data, adapting the model to the specific characteristics of the target domain. In the third step, we trained the expert model on source data in the target style. By aligning the feature distributions between the source and target, our method enables the model to generalize effectively across diverse datasets. The unsupervised training increases the computation time only during training not in inference. We evaluate the pipeline on both character-level and historical handwritten line recognition tasks, demonstrating its flexibility. We establish a new state-of-the-art on the Safran-MNIST-DLS dataset defined in Track 2 of the DAGECC competition with 76.17% Macro F1-score and achieve competitive results on each of the specific datasets of READ 2018 without using the labels of the target data. Code is available at this https URL
- [266] arXiv:2611.17953 [pdf, other]
-
Title: The Transformer Paradigm: Dynamic Processing and a New Principle of InferenceComments: v1: 15 pp., 5 figs., 1 tableSubjects: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR); Computers and Society (cs.CY); Emerging Technologies (cs.ET); Neural and Evolutionary Computing (cs.NE)
We argue that the Transformer revolution is not adequately understood as a merely quantitative expansion of deep learning, but as the emergence of a new paradigm based on a different principle of inference. We interpret the Transformer as a system in which token representations are transformed by parameterized transformations, or transforming concepts. During inference, some of these transformations are dynamically constructed from the current input: previously acquired transformations are first adapted and then applied. At the mechanical-structural level, this dynamic processing is realized through output-weight interconnections, whereby outputs of some networks determine the weights of others. The resulting mode of processing, which we call Sequence-level Interactive Dynamic Parallel Processing (SIDPP), provides a framework for explaining strong prompt sensitivity and opens research directions concerning interpretability, control, alternative AI architectures, neuromorphic systems, and possible biological realizations of dynamic processing.
- [267] arXiv:2611.17954 [pdf, html, other]
-
Title: In the Eye of the Beholder: Human Label Variation in Textual and Visual EntailmentComments: 22 pages, 18 figuresSubjects: Computation and Language (cs.CL)
Human Label Variation (HLV) is increasingly recognised as an informative signal that reflects meaningful differences in how people interpret the same input. While HLV has been studied extensively in Natural Language Inference (NLI), much less is known about HLV in Visual Entailment (VE; Xie et al., 2019), where the textual premise is replaced by an image, keeping the hypothesis and label space fixed. We introduce VariNLI-VE, a dataset for studying how HLV changes between paired NLI and VE instances. We re-annotate a subset of originally neutral SNLI (Bowman et al., 2015) instances and their visual counterparts, with the same annotators evaluating both the textual and visual versions. Each annotation includes a label, confidence score, and explanation, yielding 2,250 annotations. Labels and explanations are collected jointly, providing ecologically valid explanations. Because the image may reveal information not stated in the text, one might expect disagreement to decrease in VE. Instead, we find no clear overall reduction: disagreement decreases for some instances while it increases for others. Explanations reveal further variation, as annotators can retain the same label while changing their rationale. When evaluating vision-language models (VLMs) on the paired data, we find lower accuracy, weaker alignment with human label distributions, and generally higher uncertainty in VE than in NLI.
- [268] arXiv:2611.17955 [pdf, html, other]
-
Title: WebArena-Pro: A Heterogeneous, Multimodal, Reproducible Benchmark for Web AgentsImene Kerboua, Fatemeh Pesaran Zadeh, Xing Han Lù, Weijian Qi, Alexander Miller, Junyi Song, Yunjia Tian, Dongjin Kang, Seyeon Choi, Marzia Nouri, Ewen Gueguen, Matteo Boglioni, Fengyuan Liu, Zeyi Liao, Mengqi Yuan, Yue Li, Alexandre Lacoste, Alexandre Drouin, Spandana Gella, Huan Sun, Gunhee Kim, Siva ReddyComments: Accepted at NeurIPS 2026 E&D TrackSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
Web agents powered by large language and vision-language models are increasingly applied to realistic browser work that spans heterogeneous applications, multimodal content, and stateful workflows. However, existing reproducible web-agent benchmarks cover only a small number of web applications drawn from a few software categories, and restrict modality to text and vision. Live benchmarks broaden site coverage but sacrifice reproducibility, since pages and data drift between runs. Moreover, existing benchmarks do not meaningfully evaluate whether agents can understand and use audio and video content embedded within web tasks. To address these gaps, we introduce WebArena-Pro, a benchmark comprising 300 tasks across 20 self-hosted web applications in six domain categories, spanning distinct interface conventions, workflows, and data models. Across the evaluated agents, the best performance is achieved by both Codex with GPT 6.1 Sol and Claude Code with Claude Opus 5.5 at 62.3%, followed by OpenCode with Kimi 3.1 Pro at 60.7%. The strongest GenericAgent configurations, Gemini 3.1 Pro omni and Gemma4 31B, both achieve 30.3%. Among reproducible, human-curated web agent benchmarks, WebArena-Pro provides the broadest application coverage and the most comprehensive multimodal support to date. The benchmark treats audio and video as core observations alongside text and vision, with dedicated actions for extracting information from each. WebArena-Pro runs each task in isolation and supports reproducible, parallel evaluation. Tasks are authored through a dedicated annotator interface, filtered by LLM-assisted triage, and finally validated by humans before release. Project page: this https URL
- [269] arXiv:2611.17956 [pdf, html, other]
-
Title: nnAEU: a Simple and Efficient CNN-Autoencoder for Hyperspectral UnmixingSubjects: Computer Vision and Pattern Recognition (cs.CV)
Deep learning-based hyperspectral unmixing (HSU) has followed a trajectory of increasing architectural complexity, from simple autoencoders to transformers and, most recently, Mamba-based models, each claiming to advance the state of the art. Contrary to this trend, we propose a lightweight CNN-based autoencoder, coined "Not New Autoencoder Unmixing" (nnAEU). It is derived from the backbone of UNMamba (2025) through a pruning study that removes the Mamba component entirely. nnAEU uses less than 10 times fewer parameters than UNMamba, and far fewer than other transformer- and GAN-based methods, while outperforming them in accuracy. We provide a complete, minimal, reproducible code, available at this https URL.
- [270] arXiv:2611.17957 [pdf, html, other]
-
Title: MacroGS: Macro-Pixel Shared Gaussian Primitives for Arbitrary-Scale Light Field Super-ResolutioComments: 11 pages, 5 figuresSubjects: Computer Vision and Pattern Recognition (cs.CV)
Arbitrary-scale light field spatial super-resolution reconstructs high-resolution views at any requested scale from low-resolution multi-view observations. Gaussian primitives enable continuous reconstruction, but generating separate primitive sets for each view redundantly encodes shared scene structure across views and causes the primitive count grow linearly with the number of views. To avoid such redundancy, we propose MacroGS, a feed-forward decoder that jointly predicts a single shared Gaussian set per macro pixel rather than generating separate sets for each view. Disparity-dependent replacement and angle-conditioned appearance allow each set to be reused across all output views, keeping the predicted primitive count independent of the number of views at each requested scale. The decoder iteratively predicts per-primitive disparities, retrieves the corresponding multi-view features, and updates the primitive states. Gaussian splatting and residual refinement then reconstruct the output light field. Despite being learned without geometric supervision, the predicted disparities strongly correlate with underlying scene geometry. Experiments across five benchmarks demonstrate that MacroGS integrates effectively with diverse light field encoders, improving reconstruction quality at fixed scales while achieving leading performance across most evaluated arbitrary-scale settings. The decoder also achieves speedups of approximately $2$-$10\times$ over state-of-the-art implicit light field decoders and $10$-$18\times$ over the view-wise Gaussian baseline.
- [271] arXiv:2611.17958 [pdf, html, other]
-
Title: Reasoning Models Confabulate to Defend Injected Beliefs About Their Own RequestsComments: 20 pages, 5 figures. Also available on Zenodo: this https URLSubjects: Machine Learning (cs.LG)
Chain-of-thought (CoT) monitoring and self-report are proposed as levers for AI oversight, on the assumption that a model's account of what it is doing is anchored to what it was asked. We show this anchor can be cut. An activation edit invisible in the prompt text, the injected forming-trace of an off-topic concept, overwrites a reasoning model's belief about its own request: across six models spanning scale, architecture, and recipe, the injected topic becomes the task it reasons about, reports back, and answers: the forced quote-back returns the injected topic in 47 of 48 model-concept cells, and on the four models with clean final haiku the injected answer follows in 99% of generations. Warnings barely move it: even one naming the true topic yields under 5% true-topic answers. Handed the truth in clean tokens, the model confabulates a story about the user ("maybe they got confused and corrected themselves") and answers with the injection anyway, its behavior and chain-of-thought corrupted together: a concrete limit on CoT monitorability. Opening the model up offers no reprieve: it attends to the true request even harder after edits, yet that attention is causally inert, and the injected content rides intact task circuits, so attention maps and circuit attributions read as on-task too; the same edit redirects the request's language and format, not only its topic. What the injection reaches is the model's reference to its own input; what it elicits is its defense of it.
- [272] arXiv:2611.17959 [pdf, other]
-
Title: DeltaVisionX: A Low-Cost Vision-Guided Automation System Using YOLOv11 and PLC-Based ActuationLara Sobh (AUST), Marc Nahhas (AUST), Francisco Katach El Halabi (AUST), Michel Owayjan (AUST, IREENA, Nantes Univ), Fadi Elias SleimanJournal-ref: 2026 12th International Conference on Mechatronics and Robotics Engineering (ICMRE), Mar 2026, Oldenburg, France. pp.323-329Subjects: Machine Learning (cs.LG)
This paper presents DeltaVisionX, a low-cost industrial automation system that integrates real-time computer vision with PLC-based actuation for object classification and sorting. The system uses the YOLOv11 (You Only Look Once) deep learning model to detect objects on an operating conveyor belt and communicates detection results to a Delta DVP-SE PLC, which actuates pneumatic pistons for sorting. A custom dataset was created, labeled, and augmented to train the YOLOv11 model, achieving a mean average precision of 0.97 with real-time inference of 100 ms per frame. Encoder-based tracking ensures reliable synchronization between detection and actuation, supporting a throughput of 80 objects per minute. Results show that affordable, locally developed automation can reduce labor dependency and improve manufacturing efficiency. The novelty of the proposed system lies in its encodersynchronized vision-PLC integration rather than in the detection model itself. Future work includes integrating a highspeed Delta robot with EtherCAT-based servo control for advanced pick-and-place operations.
- [273] arXiv:2611.17963 [pdf, html, other]
-
Title: Toward Universal Translatability: On the Statistical Ontology of Generative AISubjects: Computation and Language (cs.CL)
This paper argues that the problem of translation functions both as a point of origin and as a normative horizon for contemporary generative AI. In current discourse around ''world models,'' their attainment is framed as the achievement of internal consistency, cross-modal alignment, and physical grounding. Taken together, these demands converge on an ideal of the model as a universal translation machine capable of establishing lossless correspondences across linguistic, visual, and physical representation spaces. I show that this ideal presupposes a metaphysical commitment that I call a statistical ontology: the belief that reality has an underlying, computable statistical structure that can be recovered through sufficient data and modeling, rather than being shaped and accessed through situated human interpretation and labor. Tracing this commitment back to Warren Weaver's 1949 Memorandum on machine translation, I outline how translation was historically recast as code deciphering and, later, as the alignment of probability distributions across structured outputs. Finally, I contrast this paradigm with approaches in Translation Studies that foreground opacity, power asymmetries, and the translator's positionality. I argue that generative AI's statistical ontology is pre-Kantian, pre-critical and obscures the infrastructures and human labor that make machine translatability---as alignment---possible in the first place.
- [274] arXiv:2611.17966 [pdf, html, other]
-
Title: From Images to Grounded Edits: A Real-World Benchmark for Chinese Grammatical Error Correction in Handwritten EssaysSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Chinese Grammatical Error Correction (CGEC) is typically formulated as a text-to-text problem. While suitable for digital texts, this formulation falls short in real-world K-12 education, where AI-assisted corrections are expected directly on essay images, which are primarily handwritten and exhibit many irregularities. To bridge this gap, we introduce Grounded Chinese Grammatical Error Correction (G-CGEC), an image-to-grounded-edit task requiring systems to predict both corrections and their positions on essay images. We propose G-CGEC-Eval, a benchmark comprising real-world images of K-12 Chinese essays with fine-grained annotations. For evaluation, we adapt IoU-based metrics from object detection and propose a more tolerant alternative based on center-distance matching. Using this benchmark, we evaluate open-source and proprietary vision-language models under a direct single-pass setting, as well as settings with auxiliary components (OCR and character grounding) coordinated via pipelining or agentic harnessing. Results show that even top-tier proprietary models perform surprisingly poorly in the single-pass setting, while pipelining and agentic harnessing yield meaningful improvements. Notably, an error-type-based multi-agent workflow achieves the best performance, establishing it as the strongest baseline for future G-CGEC research. Through ablations and fine-grained analysis, we find the task remains far from solved, with linguistic error discovery and character-level grounding as the primary bottlenecks. This underscores the need for models that better integrate visual perception with linguistic reasoning and visual grounding.
- [275] arXiv:2611.17967 [pdf, html, other]
-
Title: Analyzing Learning-Dynamics Across Loss and Prediction Choices for Diffusion ModelsComments: NuerIPS2026Subjects: Machine Learning (cs.LG)
Denoising generative models predict clean data ($x$), noise ($\epsilon$), or velocity ($v$), while training losses can be measured in any of these spaces. Although the resulting nine combinations are algebraically related, they can yield substantially different denoising accuracy and sample quality. We disentangle the effects of prediction and loss spaces. For a fixed prediction target, we show that changing the loss space is exactly equivalent to timestep reweighting and provide a heuristic analysis of how this reweighting affects generation quality. We then show that a rank-$m$ linear bottleneck compressing a $D$-dimensional noisy input imposes an irreducible squared-error floor of $D-m$ on noise and velocity prediction, whereas clean-data prediction can exploit low-dimensional data structure. Combining this bound with prediction-dependent error conversion yields distinct necessary conditions on bottleneck capacity. Experiments with U-Net and JiT across CIFAR-10, Imagenette, and ImageNet-256 support these theoretical predictions. Under compression, prediction choice strongly affects both denoising accuracy and sample quality, with noise-prediction errors closely matching the theoretical lower bound. Our results provide a quantitative understanding of how architectural bottlenecks interact with prediction and loss choices in denoising generative models.
- [276] arXiv:2611.17968 [pdf, html, other]
-
Title: TotalSegmentator Aorta Report: Automated Aortic Diameter Measurement at 11 Standardized Landmarks on CT ImagesJakob Wasserthal, Daniel Hinck, Martin Segeroth, Joshy Cyriac, Denis Berdajs, Jules Miazza, Maurice PradellaSubjects: Artificial Intelligence (cs.AI)
Purpose: To develop and evaluate a fully automated, open-source tool that measures aortic diameters at 11 standardized landmarks from the aortic annulus to the bifurcation on any CT image, following current guideline recommendations.
Materials and Methods: This retrospective single-center study included 239 arterial phase CT angiography examinations of 230 patients (median age, 65 years [range, 18-99 years]; 163 men) performed between April 2025 and October 2026 (209 electrocardiographically [ECG] gated). The tool locates landmarks from TotalSegmentator segmentations of aortic branches and vertebrae and measures the maximum diameter at each. A senior radiologist reviewed and, where needed, corrected the automated measurements to create the reference standard. Agreement was assessed with mean absolute error (MAE), bias, and differences greater than 2 mm.
Results: Across all landmarks, MAE was 1.06 mm (95% CI: 0.91, 1.22), and 86.6% (95% CI: 84.4, 88.5) of differences were 2 mm or less. The tool measured slightly larger diameters (bias, 0.45 mm [95% CI: 0.34, 0.57]). The sinuses of Valsalva had the largest MAE (2.17 mm [95% CI: 1.59, 2.86]). For 1.2% of landmarks, the automatic measurement failed. Examination-level MAE did not differ significantly by pathology, atherosclerosis grade (P=.95), or ECG gating (P=.14). Examinations with mural thrombus more often had an MAE greater than 2 mm (50% [95% CI: 31, 69] vs 20%; P=.03).
Conclusion: The tool measured aortic diameters at 11 landmarks along the entire aorta with an MAE of about 1 mm, below published interreader variability. Agreement was lower at the aortic root and with mural thrombus.
Code availability: The tool is open source and freely available at this https URL. - [277] arXiv:2611.17973 [pdf, html, other]
-
Title: Wireless Node Ranking via Fourier-Driven Maximum-Likelihood Estimation (extended version)Subjects: Information Theory (cs.IT); Networking and Internet Architecture (cs.NI)
We apply non-abelian Fourier theory to the problem of maximum-likelihood estimation of the order of a line topology, based on signal strength measurements under shadowing only. This yields up to a 60-fold speedup over standard methods for uncorrelated shadowing.
Moreover, for correlated Gaussian shadowing with a singular covariance (e.g. the joint path-loss model), maximum-likelihood recovery is exact and independent of the shadowing variance (an infinite Le Cam threshold), which our implementation confirms for $\sigma/\alpha$ as high as 7.0. The key object is the covariance's null space. Its dimension is fixed by the topology; its orientation depends on the correlation function only through the per-link variances.
Although the mechanism formally requires an exact line, it is robust to small perturbations of the formation: an approximate kernel recovers as well as the exact one. We explore the dependence on the perturbation in detail for five nodes. - [278] arXiv:2611.17975 [pdf, html, other]
-
Title: Enhancing Transient Stability of VSG-Controlled Converters Under Grid Frequency DisturbancesSubjects: Systems and Control (eess.SY)
Strong damping feedback can substantially improve the transient synchronization stability of renewable energy converters employing virtual synchronous generator (VSG) control. However, conventional damping feedback uses the nominal frequency as its reference. When the frequency of a weak grid deviates from nominal, strong damping tends to keep the VSG frequency near nominal. This can sustain a frequency difference between the VSG and the grid and compromise transient synchronization. To address this problem, this paper proposes a complex-frequency-based method to estimate the grid reference frequency from local terminal measurements. The estimated frequency replaces the nominal reference in the VSG synchronization loop, so that damping acts on the frequency difference between the VSG and the real grid. Controller hardware-in-the-loop tests validate the effectiveness of the proposed method under grid frequency disturbances and its robustness to grid parameter uncertainty and adverse grid conditions.
- [279] arXiv:2611.17980 [pdf, other]
-
Title: Algorithmic Opacity as a Principal-Agent ProblemComments: 76 pages, 1 tableJournal-ref: N.Y.U. Journal of Intellectual Property & Entertainment Law 15(2) (2026) 1-76Subjects: Computers and Society (cs.CY); Machine Learning (cs.LG)
Concerns about gaming and trade secrecy, frequently invoked to justify algorithmic opacity, can also mask self-serving strategic behavior by decision-makers, who may prioritize private benefits over social welfare. We develop a principal-agent test for when algorithmic decision-making should trigger disclosure obligations even when concerns about gaming or trade secrecy exist. The test, which is anchored in decision quality, relies on two diagnostics that can be disclosed without enabling gaming or revealing trade secrets: the error profile (false positive and false negative rates, overall and for each group) and the distribution of decision outcomes across groups. These diagnostics make visible whether decision-makers align design and implementation with the public interest or instead use opacity to obscure bias, inefficiency, or strategic shirking. Strong disparities across social groups, and asymmetries in the distribution of decision errors across outcome types, should trigger rebuttable presumptions favoring further disclosure. Where ground-truth error rates are unavailable, outcome distributions across groups can support the same presumption, and a decision-maker's inability to produce an error profile is itself diagnostic. The test thus treats fairness metrics as triggers for further disclosure and extends to generative and agentic systems through analogous metrics. The framework informs differentiated disclosure regimes (staged, partial, mediated, or audience-segmented). Once such regimes are considered, the domain of justifiable algorithmic opacity is significantly narrower than commonly assumed: when gaming or trade secrecy risks are legitimate, actionable disclosures (such as training data characteristics and key system design parameters) ordinarily can be mandated without compromising innovation or enabling gaming, even for technically inscrutable algorithms.
- [280] arXiv:2611.17981 [pdf, html, other]
-
Title: When Does Oversampling Help? A Decision Framework for Imbalanced Data Streams Under Label Delay and Concept DriftSubjects: Machine Learning (cs.LG)
For imbalanced data streams with delayed labels, is it worth adding a streaming oversampler, or does continuous re-evaluation of earlier predictions already recover the cost of the delay? We address this through an extensive experimental comparison of streaming oversampling and continuous re-evaluation under joint class imbalance, concept drift, and label delay. Streaming oversamplers are typically evaluated under the test-then-train assumption of immediate labels; their behaviour under these conditions, assessed through re-evaluation, has not been systematically studied. We close this gap with a decision framework that predicts when oversampling adds value over re-evaluation, organised around three axes: drift density, minority-class learnability, and delay relative to the refinement window, and operationalised as a competition between a recovery channel (re-evaluation) and a minority-coverage channel (oversampling). Evaluating $15$ classifiers on $6$ synthetic and $6$ real streams, we find: re-evaluation recovers delay-induced loss on drift-active streams but is silent on stationary ones; oversampling's marginal contribution over re-evaluation is statistically detectable in only one real-data configuration; and cost--performance Pareto recommendations from synthetic streams do not transfer to real data, where fronts partition by drift regime. Method selection should therefore be conditioned on the stream's profile and re-estimated as it evolves, not inherited from synthetic benchmarks.
- [281] arXiv:2611.17983 [pdf, other]
-
Title: Interactive deep-learning based interface for brain tumour segmentationJournal-ref: SPIE Medical Imaging, Feb 2026, Vancouver, Canada. pp.17Subjects: Artificial Intelligence (cs.AI)
Brain tumour segmentation from MR images is an integral part of their treatment, either through surgical resection or radiotherapy. Both applications require a high degree of precision that can be difficult to ensure for all patients due to the high degree of variability in the size, shape, number, and location of these tumours. This leads to difficulties with their automatic segmentation via deep learning, which can display catastrophic segmentation errors when presented with images that differ, even subtly, in terms of the specific MR sequence used, which varies from vendor to vendor. This article presents a deep-learning-based interactive segmentation system that, while being fully automatic by default, has the capability of being easily corrected by human users, allowing them ensure sufficiently high brain tumour segmentation accuracy for all patients. This interface is evaluated both in terms of the accuracy of its automatic approach as well as its usability through a preliminary observational expert user study.
- [282] arXiv:2611.17989 [pdf, other]
-
Title: The Exact Stabilization Time of Self-Stabilizing Coloring: Mean, Variance, and Distribution{É}lie de Panafieu, Fabien Mathieu (NPA), Maria Gradinariu Potop-Butucaru (NPA), S{é}bastien Tixeuil (NPA, IUF)Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Combinatorics (math.CO); Probability (math.PR)
The worst-case stabilization time of a self-stabilizing algorithm usually says little about its typical behavior beyond the order of magnitude. This paper revisits the classical self-stabilizing $Delta+1$ coloring algorithm with vertices activated at random, one at a time (the random central daemon). Considering the number of color changes as a random variable, we study its mean, variance and, on several families of graphs, its exact distribution. The analysis rests on a reduction: a vertex changes color, at most once, exactly when it is not a local maximum of a uniformly random labeling of the vertices, maxima being taken among the neighbors that share its initial color (the monochrome initial configuration is the slowest case). What can be predicted depends on what is known of the graph. The degree distribution (the view at distance one) gives the mean. The variance needs the view at distance two, in particular the triangles, and the full law needs the whole graph. We give closed-form results, from monochrome and from uniformly random initial configurations, on paths, cycles, random trees, Erdős-R{é}nyi graphs and bipartite random graphs. From a monochrome start, the mean stays within a factor two of the worst case, whereas the variance ranges from zero (disjoint cliques) to order $n^2$ (the star on $n$ vertices)..
- [283] arXiv:2611.17991 [pdf, other]
-
Title: LACCO: Local Accumulation with Computation--Communication OverlapSubjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Training language models across low-bandwidth networks requires reducing synchronization overhead. Local optimization methods such as DiLoCo communicate only after several local steps, but outer gradient averaging and the outer update remain on the critical path. Existing computation-communication overlap methods avoid this stall through delayed or replica-local updates, which can degrade convergence or cause replicas to drift. We propose LACCO (Local Accumulation with Computation-Communication Overlap), a two-stage algorithm that overlaps global communication with local computation while applying globally aggregated updates shared by every replica, without introducing staleness in the outer update. We establish convergence guarantees for smooth non-convex objectives. Experiments on language-model pre-training with FineWeb show that LACCO consistently outperforms communication-overlapped baselines such as Eager and Delayed DiLoCo. These gains remain robust as model size scales up to 0.5B parameters and across a range of communication intervals and numbers of replicas up to 16 workers, while incurring only negligible-or fully overlappable-computational or memory overhead, which demonstrates the strength of LACCO.
- [284] arXiv:2611.17998 [pdf, html, other]
-
Title: Understanding the limits of snapshot-based fault diagnosis in distribution networksSubjects: Artificial Intelligence (cs.AI)
Inverter-based resources weaken fault currents in distribution networks, and snapshot-based diagnosis, which classifies a fault from voltage phasors at a few buses at one instant, offers a cheap first-stage screen. Its published accuracies, however, rest on different noise assumptions and do not show whether the error comes from the classifier, the features, the measurement error, or the fault resistance. This paper separates these sources on three IEEE test feeders and a public data set with distributed generation, using four feature sets, thirteen classifiers, and measurement error defined by total vector error classes. The classifier is not the limiting factor, because the full phasor snapshot classifies every fault correctly without noise, and the remaining error comes from measurement error on high-impedance faults. Three voltage sensors classify 90-96% of faults at the 1% accuracy class and 96-99% with distribution micro phasor measurement units (micro-PMUs). A discriminability-margin placement selects them from simulations alone in under one second and matches the best of twelve placement methods. The results show when snapshot-based diagnosis can be trusted.
- [285] arXiv:2611.18003 [pdf, html, other]
-
Title: Reusing Locomotion Training Recipes Across Humanoids via Body-Aligned Virtual JointsComments: 8 pages, 7 figures, 1 tableSubjects: Robotics (cs.RO)
Reinforcement learning has enabled robust humanoid locomotion, yet training recipes (rewards, architectures, hyperparameters) remain hard to reuse across robots with different native joint coordinates, since policies output native joint angles while action scales, bounds, gains, and penalties are defined in those robot-specific coordinates. Existing transmission models convert native targets into actuator commands but do not standardize the coordinate interface exposed to the policy. We introduce a body-aligned virtual joint space upstream of this mapping: policy actions and observations follow a common anatomical convention, and a deterministic kinematic layer maps virtual targets to native ones, leaving the existing transmission and low-level control stack unchanged. Under a single unmodified recipe on three humanoids with different leg-axis layouts, it is the only interface that tracks accurately, keeps the feet aligned with the travel direction, and trains stably on all three, and on Unitree G1 hardware it reduces heading drift by nearly a factor of six.
- [286] arXiv:2611.18005 [pdf, html, other]
-
Title: RiverScan: Learning to Accumulate on River NetworksComments: 17 pages including references and appendices; 6 figures, 5 tablesSubjects: Machine Learning (cs.LG)
Accurate runoff prediction across large river basins requires models to capture heterogeneous local hydrological responses and their accumulation downstream. River-network connectivity provides a natural structure for this task, but using it effectively requires learning how to combine contributions from subcatchments of different sizes. We frame this challenge as learning to accumulate and introduce RiverScan, a spatial operator based on area-weighted selective state-space recurrence. RiverScan recursively combines local temporal representations with accumulated upstream states in topological order. Contributing-area fractions account for subcatchment size, while learned retention controls how upstream information persists downstream. Experiments on 608 stations in the Danube basin show that RiverScan outperforms the evaluated temporal and graph baselines in overall predictive performance. Structural controls and ablations demonstrate that composing heterogeneous subcatchment representations in a manner aligned with hydrological accumulation substantially improves runoff prediction, with the benefits of correct topology growing along the river-network hierarchy. These findings establish physically structured accumulation as an effective inductive bias for hydrological modeling and motivate finer-grained, fully distributed AI-based hydrological models.
- [287] arXiv:2611.18006 [pdf, html, other]
-
Title: RecurIR: Learning When and How to Revise Restoration Policies via Recursive Self-ImprovementSubjects: Computer Vision and Pattern Recognition (cs.CV)
Agent-based image restoration has achieved promising performance by coordinating restoration tools to handle heterogeneous and coupled degradations. However, existing agents often improve their decisions through accumulated successful and failed executions, without explicitly identifying the conditions under which learned restoration strategies remain effective. Consequently, changes in degradation severity, spatial coverage, or interactions may invalidate previous decisions, while providing limited guidance on how these decisions should be revised. In this paper, we propose \textbf{RecurIR}, a recursive self-improvement framework that enables restoration agents to learn when and how to revise their decision rules. Specifically, we introduce a policy-hypothesis discovery mechanism to identify testable, condition-dependent rules through matched tool executions. We then design a rule-falsification-driven task evolution mechanism to generate targeted degradation variations, examine the validity of learned rules, and identify their applicability boundaries. Based on the resulting evidence, we develop a change-conditioned policy revision mechanism that learns how variations in degradation conditions should modify tool selection, execution order, operation strength, and processing regions. We further incorporate independent verification to retain effective revisions and use the resulting feedback to guide subsequent hypothesis discovery and task evolution. Through this recursive process, RecurIR progressively refines its restoration policies as degradation conditions change. Extensive experiments demonstrate the effectiveness of the proposed framework.
- [288] arXiv:2611.18008 [pdf, html, other]
-
Title: Sample-Efficient Control of Magnetoconvective Flows via Differentiable SimulationSubjects: Machine Learning (cs.LG); Fluid Dynamics (physics.flu-dyn)
Magnetohydrodynamic (MHD) fluid flows are central to demanding engineering applications, most prominently fusion blanket cooling, where flow control is a promising tool to optimize key properties such as heat transfer. While active flow control (AFC) has recently advanced through machine learning, its application to MHD flows remains largely unexplored. In this work, we present $\phi$-PICT, a novel differentiable GPU-accelerated MHD solver targeted at AFC. We validate $\phi$-PICT against MHD benchmarks and demonstrate its high computational efficiency. Integrated into FluidGym, $\phi$-PICT enables sample-efficient gradient-based algorithms to improve heat transfer by $>25$% and allows optimizers to achieve $10^3\times$ higher accuracy in inverse design than black-box methods. $\phi$-PICT is publicly available at this https URL.
- [289] arXiv:2611.18014 [pdf, html, other]
-
Title: Conflict, Feasibility, Faithfulness, Displacement: Four Quantities That Come Apart in Federated Dual-Cone AggregationSubjects: Machine Learning (cs.LG)
FedFV, FedGH and FedMGDA+ aggregate client updates with multi-objective gradient aggregators to remove conflict between them. We examine this rationale with UPGrad, which increases no client's loss to first order, in a comparison that changes only the direction of FedAvg's update. Four quantities the rationale treats as one -- conflict between client updates, feasibility of their weighted mean, faithfulness of the transmitted updates to the clients' gradients, and displacement of the update -- come apart. Under Dirichlet partitioning UPGrad improves CIFAR-10 accuracy by +0.78 pp (14/15 seeds, p < 0.001), only with full participation; on Oxford-IIIT Pet its +0.52 pp gain vanishes (+0.07 pp) once the data our protocol left unused are kept. Varying only the Dirichlet concentration, the gain grows with heterogeneity, yet the split with the most conflict shows none. Uniform client weights, blind to conflict, match the CIFAR-10 gain and exceed the Pets gain; a matched random rotation does neither. No fairness effect survives multiplicity correction, and on CIFAR-10 the guarantee does not reach the clients' gradients. We state what evaluations should report.
- [290] arXiv:2611.18017 [pdf, html, other]
-
Title: On Structure-Aware SAT SolvingSubjects: Logic in Computer Science (cs.LO)
Although modern SAT solvers have made remarkable progress, they still exploit formula structure only to a limited extent. We present two techniques that make SAT solving more structure-aware and provide experimental results showing their promise. The first technique ``$\mathit{melts}$'' the formula by reducing interaction between its parts. This is achieved by adding new variables, using existential quantifiers, and applying a procedure called partial quantifier elimination. The second technique exploits the notion of a $\mathit{local}$ $\mathit{proof}$. The latter establishes that a formula is unsatisfiable by proving that certain literals of its clauses are redundant. We show that using local proofs makes SAT solving more formula-specific.
- [291] arXiv:2611.18020 [pdf, html, other]
-
Title: Data-Driven Output Regulation via Predictive ControlSubjects: Systems and Control (eess.SY)
This paper addresses constrained output regulation for unknown linear systems by means of predictive control directly from input-output data. The objective is to reject disturbances and track references generated by a known exosystem, without identifying a plant model or solving the output regulator equations. The first step is a trajectory parametrization for plants driven by exosignals with known dynamics. % By exploiting the known exosystem eigenstructure, Willems et al's fundamental lemma is extended from autonomous input-output trajectories to trajectories coupled with exogenous modes. By exploiting the known exosystem eigenstructure, input-output trajectories driven by an exosignal of known phase are parametrized directly from data, which contains the affine version of Willems et al.'s fundamental lemma as a special case. This parametrization yields a nominal data-driven predictive control (DDPC) regulator, for which a Lyapunov argument proves exponential convergence to the regulator manifold. For periodic exosignals, the steady-state feedforward input is eliminated from the online problem through an incremental-input regularization, leading to an implementable regulator based on measured data and known exosystem information. Bounded output noise is handled by slack and coefficient regularization, with a regularization center computed offline from data; the resulting robust DDPC law is recursively feasible and practically stabilizing, with an ultimate regulation error that vanishes with the noise bound. A constrained lateral vehicle-control example with periodic road curvature illustrates disturbance rejection, input-constraint satisfaction, and the transition from exact regulation with noise-free data to practical regulation with noisy data.
- [292] arXiv:2611.18023 [pdf, html, other]
-
Title: RSSA: Robust Semantic and Spatial Aligner for Collaborative PerceptionComments: Accepted to NeurIPS 2026. 18 pages, 14 figures, 11 tables. Code: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV)
V2X collaborative perception improves single-vehicle perception by aggregating multi-view features from collaborative agents. However, existing methods typically rely on simple spatial warping to align the collaborative features, which can cause semantic misalignment during fusion. Meanwhile, these deterministic warping-based methods are highly sensitive to the positional noises. In this paper, we propose a plug-in module called Robust Semantic and Spatial Aligner (RSSA), which performs semantic and probabilistic spatial transformations for robust feature alignment. Specifically, a Semantic Transformation (SemT) block modulates feature channels based on the observed relative pose to correct semantic inconsistencies. Subsequently, a Probabilistic Spatial Transformation (P-SpaT) block performs probabilistic warping by sampling from the Gaussian-modeled relative pose space. A Relative Transformation Augmentation (RT-Aug) strategy is further introduced to augment the relative poses and facilitate the training of the RSSA. Extensive experiments on both simulated and real-world datasets demonstrate that RSSA improves the state-of-the-art methods by a large margin under the configuration of both clean and noisy positions.
- [293] arXiv:2611.18024 [pdf, html, other]
-
Title: JaRS: Contrast-Free Localization of Memorized Regions in Diffusion Models via Jacobian Row SumsSubjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Diffusion models sometimes memorize and reproduce regions of their training images, raising privacy and copyright concerns. We aim to localize these regions. Existing contrast-based methods use the diagonal of the denoiser Jacobian, which reflects the posterior variance of each pixel. However, a small diagonal entry is ambiguous: it signals either memorization or averaging over many neighboring pixels in smooth or predictable regions. To resolve this ambiguity, these methods subtract a reference map, which requires a second evaluation with a null prompt or an earlier checkpoint. Unconditional models have no null prompt, and earlier checkpoints are unavailable in general. We show that Jacobian row sums (JaRS), unlike the diagonal, separate memorization from local averaging without a reference map. Pairing JaRS with an image-level detection score that uses the same statistic at an early denoising step yields JaRS+. Across Stable Diffusion v1.4 and v2.1, Realistic Vision v5.1, and an unconditional diffusion Transformer (DiT), JaRS+ outperforms contrast-based localization methods.
- [294] arXiv:2611.18026 [pdf, html, other]
-
Title: A Generalizable Approach to Safety-Aware Robotic Behavior Planning with an Explicit Constraint SpaceXianxu Wu, Shang-Ching Liu, Zhaocong Liang, Hengxiang Chen, Yujian Ma, Shenwen Deng, Qiang Li, Jianwei ZhangSubjects: Robotics (cs.RO); Machine Learning (cs.LG)
Large language and vision-language models have substantially advanced robotic behavior planning from natural-language instructions and visual observations. However, safe manipulation in tabletop environments remains challenging because the safety constraints required for a task depend on the surrounding scene and may not be explicitly specified in the instruction. We propose a safety-aware behavior planning framework based on an explicit Constraint Space that stores reusable textual safety rules. Rather than retrieving rules solely by semantic similarity, we formulate constraint selection as task- and environment-conditioned applicability learning. A lightweight query-conditioned dual-tower model identifies the constraints applicable to the current context, which are then provided together with the original task and image to a multimodal behavior planner. To handle incorrect constraint selections and previously unseen hazards, we introduce a multimodal Safety Inspector that evaluates generated primitive plans, reconstructs the complete applicable constraint set, and provides corrective planning feedback. Validated inspection feedback is further used for test-time adaptation of constraint applicability and parameter-efficient adaptation of the Safety Inspector through LoRA, while newly identified constraints are incorporated into the Constraint Space. The resulting framework integrates explicit safety knowledge representation,context-aware constraint selection, corrective safety inspection, and test-time knowledge adaptation for safety-aware robotic behavior planning.
- [295] arXiv:2611.18027 [pdf, html, other]
-
Title: Beyond Bot Labels: Retrospective Consequence-Grounded Ranking of Attributed Cache Harm in E-Commerce LogsComments: 11 pages, 4 figures, 7 tables. Accepted to appear in Proc. of the 2026 IEEE International Conference on Big Data (IEEE BigData 2026), Phoenix, AZ, USA. This is the author's accepted manuscriptSubjects: Cryptography and Security (cs.CR); Machine Learning (cs.LG); Networking and Internet Architecture (cs.NI)
Web caches reduce origin load and latency, but automated traffic can either warm them or evict objects shoppers reuse, lowering the cache hit ratio (CHR). Existing bot defenses generally infer automation from User-Agent (UA) strings or behavior rather than identify sessions that impose cross-session cache harm. We instead formulate the task as ranking sessions by eviction damage and propose consequence-grounded session ranking. A single chronological replay attributes each eviction-induced miss by a protected shopping session (a victim) to the session whose insertion last evicted the object. A behavioral model predicts the target from session behavior alone. Selected sessions bypass the cache but remain origin-served. In a controlled check of 539 non-victim sessions selected across harm levels, attribution order tracks victim-hit changes after removal (Spearman correlation 0.80). On a 12.7-million-request real trace, the top 1% of sessions carry 48.8% of attributed harm. Sessions without bot tags contribute 49.1% of this tier's harm. At the 20% request-volume target, the default conditional CHR difference reaches 93% of the UA baseline, which selects 47 times as many sessions. Six training seeds show little variation. This comparison conditions on the victims that remain, so we also report victim bypass. Local replay and retraining on a second store reproduce harm concentration and a top tier dominated by sessions without bot tags. These results establish attributed cross-session harm as an outcome-based target for retrospective cache audits and selective intervention design.
- [296] arXiv:2611.18028 [pdf, html, other]
-
Title: Long rational and Gaussian integer multiplicationComments: 38 pagesSubjects: Data Structures and Algorithms (cs.DS); Number Theory (math.NT)
Long integer multiplication is a fundamental operation in computational algebra, heavily relying on Toom--Cook and related evaluation-interpolation algorithms. Conventional wisdom dictates that multiplication on the Gaussian integers $\mathbb{Z}[i]$ is reduced to standard rational integer multiplication, since $\mathbb{Z}[i]$ is a ring extension of $\mathbb{Z}$. We show that this approach is far from optimal and moreover inverting this established paradigm, that is reducing rational to Gaussian multiplication, leads to faster rational integer multiplication algorithms.
We introduce a novel Toom--Cook framework that reduces rational long integer multiplication to Gaussian integer multiplication, yielding a strictly more efficient computational pipeline. By taking advantage of the special properties of the rational integer domain inside the Gaussian integers, we achieve a more favourable algebraic structure that reduces the overall arithmetic cost in every metric. To validate this theoretical shift, we provide a complete implementation of our algorithm. Benchmarks demonstrate that our approach outperforms the highly optimised GNU Multiple Precision Arithmetic Library (GMP) by up to 29\%, bridging a fundamental theoretical insight with significant practical improvements in high-performance computer algebra. - [297] arXiv:2611.18030 [pdf, html, other]
-
Title: PerfArena: Ranking LLMs by Code PerformanceSubjects: Performance (cs.PF); Computation and Language (cs.CL); Software Engineering (cs.SE)
Industry now relies on LLMs to write a large share of production code. Although that code is mostly, but not always, correct, correctness alone does not determine whether it is fit for repeated and extensive use, which depends on what the code costs in time, memory and energy. Most code-generation benchmarks nonetheless evaluate correctness in isolation, and most performance-oriented benchmarks cover few languages and rarely measure energy consumption. To address this gap, we present PERFARENA, an open-source, reproducible arena that ranks LLMs by the performance of their generated code across multiple languages and competitions. PERFARENA executes the solutions of each LLM on a Unix-like host and reports correctness together with latency, peak memory, and energy consumption. Correctness oracles, performance oracles, profilers, and code-generation workflows are pluggable in PERFARENA, enabling the integration of new judging backends and measurement tools without changing the pipeline. We evaluate PERFARENA on two competitions: CLBGComp, 10 compute-intensive benchmarks from the Computer Language Benchmarks Game, and LeetCodeComp, 99 LeetCode problems whose accepted solutions differ measurably in latency. Currently, PERFARENA holds 9,455 correct and measured solutions from 13 state-of-the-art LLMs across 7 programming languages. Across both competitions, got-6-astra is the only LLM that no other LLM beats significantly in latency, energy consumption, or peak memory, solving 99.7% of LeetCodeComp and 100% of CLBGComp. All PERFARENA leaderboards and measured solutions are publicly available, providing a continuously updated reference that helps developers choose an LLM per language and metric, instead of relying only on correctness rates or pricing.
- [298] arXiv:2611.18031 [pdf, html, other]
-
Title: Latent resonances in oscillator networksComments: 22 pages, 5 figures, Supplementary Materials included (6 figures, 7 tables)Subjects: Systems and Control (eess.SY); Adaptation and Self-Organizing Systems (nlin.AO); Physics and Society (physics.soc-ph)
In networks of coupled oscillators, from neural circuits to power grids, damping is expected to suppress oscillations because of its dissipativity. Here we show that local damping can, counterintuitively, amplify oscillations along some input-output channels, while the network's global oscillation energy decreases and its spectral properties improve. The amplified oscillation arises from a latent resonance that lies in the spectrum yet barely manifests. The added passive damping reshapes this resonance by mixing it with adjacent eigenmodes. In random power-grid networks, damping can boost weakly responding channels by more than an order of magnitude, while subdominant channels can grow to exceed the originally strongest channel. Damping thus not only suppresses oscillations but also provides a way to redistribute oscillatory energy. For any passive network, the amplified response and its upper bound follow exactly from baseline quantities, offering a principled route for network design and control.
- [299] arXiv:2611.18033 [pdf, html, other]
-
Title: Multi-level latent-diffusion adaptation for cross-region point cloud segmentation of ancient terracesComments: 33 pages, 6 figuresSubjects: Computer Vision and Pattern Recognition (cs.CV)
Automated lidar feature detection from lidar point clouds can support archaeological surveys, but models trained in one landscape may transfer poorly to another because of differences in terrain, vegetation, acquisition characteristics, and feature morphology. We address this problem for cross-region detection of ancient terraces using transductive unsupervised domain adaptation that exploits an unlabelled target survey. We propose MELD (Multi-level Embedding-space Latent Diffusion), which trains separate diffusion denoisers on target-domain representations at four hierarchical PointNet++ feature levels. These target distribution priors are then used to refine source representations during supervised training and target representations during inference, allowing adaptation to act directly on the multi-scale features consumed by the segmentation decoder. We additionally investigate local geometric descriptors derived from K-nearest-neighbour neighbourhoods, including surface normals, slope, and aspect. MELD is evaluated bidirectionally between the Teyuna and Miraflores lidar surveys in the Sierra Nevada de Santa Marta, Colombia, against a source-only model and nine configurations of established domain adaptation approaches. MELD raises F1 over the source-only model from 0.397 to 0.459 for Teyuna-to-Miraflores transfer and from 0.557 to 0.594 in the reverse direction, achieving the highest F1 among the evaluated adaptation methods in both directions. Adding local geometry further raises F1 to 0.542 and 0.606, respectively. Multi-level ablation shows that refining all four PointNet++ levels provides the strongest performance in both directions. These results show that explicit modelling of target feature distributions in hierarchical point cloud feature space, complemented by local geometry, provides an effective route to cross-region archaeological mapping without target labels.
- [300] arXiv:2611.18034 [pdf, other]
-
Title: From institutional journalism to political extremes: Structural asymmetries in YouTube's news recommender during the 2022 French presidential campaignJulien Figeac (UTOPI), Erwan Le Merrer (ARTISHAU), Marie Neihouser (CESSP), Gilles Tr{é}dan (LAAS-TRUST)Subjects: Social and Information Networks (cs.SI)
Debate has intensified over how algorithmic recommendations shape news visibility, particularly regarding structural and ideological bias. This study examines YouTube's News Recommender System (NRS) in the French ``News'' section, using a large-scale audit based on automated browsing agents operating under cold-start conditions, without prior viewing history, to isolate the algorithm's structural baseline independently of user preferences. Our results reveal that YouTube's NRS operates within a structurally asymmetric visibility regime: even in the absence of prior user behaviour, recommendation outputs disproportionately privilege platform-native creators and politically engaged content over established journalism. Television channels retain some relative advantage, but institutional news outlets, particularly press and radio, are systematically marginalised across recommendation sequences. Among political content, figures affiliated with far-left and far-right movements occupy a disproportionately visible position in recommendation outputs. Rather than correcting this asymmetry, continuity effects reproduce and stabilise it across recommendation sequences. These results show that visibility in YouTube's News section cannot be reduced to accumulated user preferences alone, but is also shaped by a structural bias embedded in the platform's recommendation system, which plays an important role in sustaining the prominence of political actors affiliated with the far left and far right.
- [301] arXiv:2611.18035 [pdf, html, other]
-
Title: RL$^{\mathrm{RT}}$: On-Device Reinforcement Learning RuntimeComments: 8 pages, 11 figuresSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
On-device reinforcement learning runs a periodic actor and a learner on one embedded GPU shared with other workloads. Because the learner produces the actor's weights, existing designs stall the actor or let its policy drift. We present RL$^{\mathrm{RT}}$, an on-device runtime that makes this coupling a tunable parameter: a refresh bound limits actor releases between refreshes, and a GPU permit with slack-aware admission prioritises the actor and fills spare time with training atop any base scheduler. Under the real refresh cost on two Jetson boards, RL$^{\mathrm{RT}}$ keeps a 16.8 M-parameter policy's deadline misses below 1%, against 85 to 100% without it.
- [302] arXiv:2611.18036 [pdf, html, other]
-
Title: C-to-Rust Translation Validation at the LLVM IR Level via AI-Assisted Refinement ProofComments: 26 pages, 10 figures, 3 tablesSubjects: Programming Languages (cs.PL)
Migrating C codebases to Rust is a promising approach to securing legacy software, but existing translators do not guarantee that a translated Rust program preserves the behavior of the C program. Neither rule-based nor large language model (LLM)-based translators have been formally verified, and existing approaches to checking translation correctness are limited to testing and bounded checks, which cannot reason about all possible executions. In this work, we present R2C2, which validates C-to-Rust translation by proving in Rocq that the Rust program refines the C program, i.e., every behavior on the Rust side is also a behavior on the C side. Since no formal semantics defines the behaviors of C and Rust programs in the same terms, we prove refinement between the two programs compiled to the LLVM intermediate representation (LLVM IR). In addition, to resolve the discrepancies between the data representations of C and Rust, we introduce a common abstraction that relates two corresponding representations through an abstract value, and specify the behavior of each library function over abstract values. We construct the refinement proof as a chain of refinement relations, where only one relation is program-specific and proved at validation time, while the others are proved for all programs in advance. Proof construction is automated by AI agents, and since Rocq checks the proofs, the agents need not be trusted. We evaluate R2C2 on LLM translations of 519 C programs drawn from a benchmark suite used in prior work on C-to-Rust translation validation. R2C2 proves the translations of 486 programs correct without bounding their executions and discovers three incorrect translations, providing valid counterexamples that existing translation validation tools cannot find. We also conduct a case study on a small real-world image compression library and demonstrate how R2C2 can handle file operations.
- [303] arXiv:2611.18043 [pdf, html, other]
-
Title: Enrichment-Oriented Virtual Screening of Perovskite Solar Cells through Reinforcement LearningComments: 55 pages, 9 figuresSubjects: Machine Learning (cs.LG)
Experimental optimization of perovskite solar cells is constrained by combinatorial diversity in absorber compositions, device layers, and processing, which makes virtual screening attractive for prioritizing recipes before fabrication. We develop a surrogate-guided reinforcement-learning framework that casts recipe generation as virtual screening, enriching candidate shortlists near high-performing experimental neighbors. A frozen Tabular Prior-data Fitted Network (TabPFN) guides recipe generation, while a reference pool of 890 measured devices enables neighborhood-based assessment. Five-fold evaluation on 1,846 devices shows top-ten selection raises mean power conversion efficiency (PCE) from 19.68\% to 23.74\%. Three learned policies are compared with six baselines under 3,000 surrogate evaluations per run. In top-100 shortlists, Proximal policy optimization (PPO) and recurrent PPO (RPPO) achieve neighborhood-supported hit rates of 23\% and 24\% (7.9- and 8.3-fold Pure Random's 2.9\%), versus 10\% (3.4-fold) for the baselines with the highest supported hit rate (L-SHADE and Ax BO, tied) at $K = 100$. Gains coexist with weak-support fractions. Of 9,000 recipes, 65.4\% match a composition family at similarity 0.80, falling to 11.7\% at 0.90. Results support enrichment-oriented prioritization but not experimental performance, seed robustness, or matched-budget advantage.
- [304] arXiv:2611.18050 [pdf, html, other]
-
Title: FEND: Far-End Detection in Outbound Voice AI CallsComments: Submitted to EACL 2027Subjects: Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
When a conversational Voice AI agent places an outbound call, the far end may be a live person, a voicemail greeting, an interactive voice response (IVR) system, or increasingly another AI system answering on the callee's behalf. Standard Answering Machine Detection (AMD) reports only whether a human or a machine picked up, a design meant for human call centres, where one machine is as uninteresting as another. That single bit does not tell a Voice AI agent what to do, since leaving a voicemail message, navigating an IVR and holding a short exchange with another AI are three different actions and a two-valued label cannot distinguish them. In 1,000 randomly sampled calls from our own outbound deployment, a live person answered 48.1%, a voicemail system 44.6%, an AI 6.9% and an IVR 0.4%. Our proposed framework, FEND, reads the far end's audio and its streaming transcript and runs one detector for each automated class (voicemail, IVR, AI callee), with human the default when nothing fires, returning a single label within 20 seconds. On a public benchmark the voicemail components reach 100% accuracy without ever misclassifying a live human, where Asterisk's AMD misclassifies 30 to 66% of them. For the AI callee, acoustic features do not transfer across recording conditions, but speech timing does, separating AI from human at 0.952 ROC-AUC, because synthesised speech leaves far shorter gaps between words. On production traffic, separating a live person from any automated far end reaches F1 0.979, and only 2.9% of live people are given an automated label, the error that matters most for an agent acting on the result.
- [305] arXiv:2611.18053 [pdf, html, other]
-
Title: InfiHall-Bench: A Benchmark for Hallucination Evaluation in Autoregressive Text-to-Image GenerationSubjects: Computer Vision and Pattern Recognition (cs.CV)
Hallucination benchmarks for text-to-image generation are predominantly designed around diffusion models, limiting their diagnostic value for visual autoregressive (VAR) models and the development of mitigation methods. We identify two limitations in existing prompt construction paradigms: many prompts are no longer challenging for modern generators, while category-based construction relies on recurring object-attribute-relation templates and rarely tests interactions among multiple visual requirements within a single image.
We introduce InfiHall-Bench, a controlled benchmark of 411 prompts constructed by systematically composing five foundational and visually verifiable skills: attribute binding, spatial relation, action relation, part consistency, and scene consistency. Each prompt covers 4.05 annotated skills on average, enabling the evaluation of hallucinations under diverse combinations of visual requirements. Using this benchmark, we find that hallucinations in VAR models exhibit a scale-wise persistence pattern: semantic errors can emerge at different autoregressive scales and are rarely corrected by subsequent fine-scale steps.
Motivated by this observation, we propose Confidence-Masked Hallucination Mitigation (CMHM), which estimates the prediction confidence of target visual tokens at each autoregressive scale during teacher-forced fine-tuning. CMHM downweights reliably predicted tokens, allowing optimization to focus on uncertain predictions across scales. Experiments show that InfiHall-Bench distinguishes models that existing compositional benchmarks rate as comparable, while CMHM consistently outperforms uniform LoRA fine-tuning for VAR-based text-to-image generation. - [306] arXiv:2611.18055 [pdf, html, other]
-
Title: DeCompCert: A Verified Decompiler for CompCert-Generated AArch64Subjects: Programming Languages (cs.PL); Cryptography and Security (cs.CR)
DeCompCert is a Rocq-verified decompiler for AArch64 assembly generated by CompCert. It reconstructs CompCert C through nine reverse passes, combining deductively verified transformations with translation validation. Untrusted OCaml and Datalog analyses propose candidate reconstructions, while verified validators accept only candidates that satisfy the required semantic correspondence. The composed correctness theorem guarantees that every defined, safe behavior of an accepted C reconstruction is matched by the input CompCert assembly; under a safety assumption, the two programs are observationally equivalent. We evaluate DeCompCert on approximately 2.16 million lines of C. With exact function and call signatures, it reconstructs and recompiles all 3,173 eligible contexts under both unoptimized and optimized CompCert pipelines. Without retained signatures, its cumulative inference portfolio validates 2,606 of 3,164 contexts (82.4%), while the platform-plus-IDA configuration validates 2,407 contexts (76.1%). Native execution and output-quality experiments further evaluate the generated C and characterize the remaining limitations in signature recovery and source resemblance.
- [307] arXiv:2611.18057 [pdf, html, other]
-
Title: SWAM: Saliency-Weighted Asymmetric Merging for Efficient Vision TransformersSubjects: Computer Vision and Pattern Recognition (cs.CV)
Vision Transformers (ViTs) achieve strong performance on image recognition tasks, yet their quadratic self-attention complexity over input tokens creates prohibitive computational costs for practical deployment. Token reduction methods offer a practical solution by progressively reducing the token sequence during the forward pass. We propose Saliency-Weighted Asymmetric Merging (SWAM), a training-based token merging framework that extends the Bipartite Soft Matching (BSM) with two saliency-aware mechanisms. We introduce a two-step matching formulation in which cosine similarity decides candidate pairs, while a multiplicative maximum saliency penalty regulates their selection for merging. We then further use asymmetric exponential saliency-weighted aggregation, which allows more salient tokens to contribute more strongly to the merged representation. A lightweight MLP-based saliency prediction module (75k parameters per stage, 1% of backbone parameters) supervised via attention rollout is trained end-to-end jointly with a DeiT-Small backbone using a composite cross-entropy, knowledge-distillation, and saliency-alignment objective. On CIFAR-100, TinyImageNet, and ImageNet-100, under a conservative schedule that reduces tokens from 196 to 100 patch tokens, with the CLS token retained throughout, 27.5% FLOPs reduction and 1.35 times speedup, SWAM matches or exceeds baseline across the reported results. Under an aggressive schedule (196 to 70 tokens with the CLS token kept throughout, 33.8% FLOPs reduction, and 1.46 times speedup), accuracy gains are achieved on ImageNet-100 (+0.23%) with minimal losses on CIFAR-100 and TinyImageNet.
- [308] arXiv:2611.18064 [pdf, html, other]
-
Title: A$^3$: Adaptive Audio Agent Framework for Collaborative Edge ComputingSubjects: Artificial Intelligence (cs.AI)
Embedded audio agents need a nearby edge server for speed, but a server-dependent agent returns nothing when the link fails. We present A$^3$, an adaptive audio agent framework that places whole pipeline stages on the server, keeps a complete fallback agent on devices that can host one, and hedges both paths with a link-aware deadline that decides when to give up on the server. On Jetson devices and an RTX A4500 server, A$^3$ matches the server-only agent on a healthy link, answers every request in a full outage on the AGX, and beats a percentile hedge in simulated outages.
- [309] arXiv:2611.18067 [pdf, html, other]
-
Title: Agent-as-a-JudgeComments: Project page: this https URLSubjects: Computation and Language (cs.CL)
Artificial intelligence has progressed from task-specific systems toward general models spanning perception, generation, reasoning, and autonomous interaction, making reliable evaluation essential for comparison, deployment, and post-training feedback. At scale, modern evaluation pipelines increasingly rely on LLM-as-a-Judge to produce scores, preferences, and critiques; in its fixed-context setting, the judge relies on the supplied evaluand and evaluation materials. Yet a reliable verdict may require evidence beyond the supplied evaluand and context---either because critical environmental states remain unrecorded in the static execution log, or because controversial factual claims require external verification. Passive judges cannot obtain that missing evidence before deciding, especially when evaluating agent trajectories, executable artifacts, or domain-specific reports. Recent work therefore develops active evaluators that seek missing evidence and update their grounds for judgment before deciding. We bring these advances together through Agent-as-a-Judge, focusing on how evaluators construct and control checking processes within the broader LLM-based judging landscape. We define the scope through evaluation-directed checks and the use of their results, covering both structured workflows and observation-driven control. We organize the design space by multi-agent collaboration, tool integration, memory, and planning, and show how these mechanisms supply the checks and evidence on which a verdict rests. We consolidate published scores and reproduce missing entries for key agentic judges; the comparison reveals uneven agentic gains and motivates dedicated agentic evaluation. We conclude by characterizing application demands and open challenges in cost, reliability, safety, and governance that condition trusted deployment of agentic judges at scale.
- [310] arXiv:2611.18068 [pdf, html, other]
-
Title: A Realizability-Based Kinetic-Fluid Domain Decomposition Method for Single-Temperature Gas MixturesSubjects: Numerical Analysis (math.NA)
This paper introduces a domain-adaptive numerical method for multi-species kinetic models that relax toward a single-temperature and single-velocity hydrodynamic limit. To this end, we derive moment realizability matrices for this regime. In this way the ideas of Levermore, Morokoff, and Nadiga are extended to non-equilibrium mixtures in which inter-species and intra-species collisions occur at comparable rates, so that all species relax to a common velocity and a common temperature. The realizability conditions are used as an indicator to divide the computational domain, at each time step, into macroscopic and kinetic regions. In the kinetic regions, where the continuum description fails (e.g., across strong shocks, non-equilibrium interfaces, or kinetic boundary layers), the method solves the multi-species kinetic equations. Close to local thermodynamic equilibrium, the Euler equations of the mixture are solved instead. Confining the kinetic computations to the non-equilibrium regions reduces the simulation time, while the results remain close to those of the full kinetic solver. The method is discretized on Cartesian grids with high-order schemes in space and time. The kinetic equations are integrated with an Asymptotic-Preserving Implicit-Explicit (IMEX) scheme, so that a consistent discretization of the single-velocity, single-temperature Euler system of the mixture is recovered when the Knudsen number goes to zero. To the best of our knowledge, this is the first realizability-based adaptive spatial domain decomposition method for gas mixtures in this single-velocity, single-temperature setting.
- [311] arXiv:2611.18069 [pdf, html, other]
-
Title: Quantifying the Advantage of Coarser DPORComments: 32 pages, 2 figuresSubjects: Programming Languages (cs.PL)
Dynamic partial-order reduction (DPOR) has emerged as a promising approach to addressing interleaving explosion in bounded model checking of concurrent programs. Given an equivalence relation on interleavings, the model checker explores a program's executions while minimizing repeated visits to the same class; optimal model checkers avoid such repetition altogether. A coarser equivalence promises faster exploration by reducing the number of classes. This has motivated alternatives to Mazurkiewicz's trace equivalence such as reads-from and reads-value-from equivalence. This promise, however, does not always materialize in practice: while a coarser equivalence can yield exponentially fewer classes on some families of programs, it may offer little or no reduction for the program at hand.
We show how to quantify the advantage of coarser equivalences for individual programs. We focus on the problem of counting the reads-from (RF) equivalence classes induced by a given program, as well as the problem of computing the ratio of RF to Mazurkiewicz class counts under sequential consistency. We show that both problems are $\#\mathrm{P}$-hard and that the reduction ratio remains hard to approximate accurately under standard complexity assumptions.
Guided by these limits, we present CoEst, a sampling-based algorithm for estimating both the RF count and its ratio to the Mazurkiewicz count. For a fixed number of threads, each trial runs in polynomial time and yields unbiased count estimates. As the number of independent trials grows, the count and ratio estimates converge almost surely to their true values. Our algorithm can provide statistical evidence for deciding whether Mazurkiewicz or RF model checking is likely to be faster for a given program. In experiments across 34 benchmark configurations, our method selects the faster checker in 99.4% of runs using a one-minute estimation budget. - [312] arXiv:2611.18072 [pdf, html, other]
-
Title: Measuring Public Participation in Local Government: An Audio-Based Instrument for Classifying Speakers in City Council MeetingsSubjects: Computers and Society (cs.CY)
How much does the public actually speak in local government? Despite sustained interest in democratic participation, we lack large-scale measures of public voice at the local level, because the underlying record -- audio of open meetings -- has been prohibitively costly to convert into data. We contribute a measurement instrument: an audio-analysis pipeline that classifies each speaker in a city council meeting as an elected official or a member of the public directly from audio, and aggregates these labels into validated measures of participation such as floor-time share. The pipeline first segments each recording by speaker, then classifies each speaker with a multimodal large language model (MLLM), using the meeting's own temporal structure (neighboring speakers, iterative refinement, and a majority vote across three classifiers) as supporting components. We collect 1,618 recordings from seven U.S. cities (2017-2023); after diarization and quality control, the analytic corpus is 1,565 meetings from five cities. The task is non-trivial: audio quality is uneven, members of the public speak informally and are often introduced only briefly, and speech frequently overlaps. Even so, segmentation reaches 89% accuracy (mean pairwise kappa=0.84), and the classifier reaches 87.7% accuracy (kappa=0.736, public-participant F1=0.832) on ground truth from five cities (N=1,761 speakers) -- agreement comparable to that among the human annotators themselves. At scale, members of the public receive a median of just 10% of floor time, the top speaker is an official in over 94% of meetings, and public participation declined over COVID-19 without recovering by 2023. The instrument measures participation directly from audio, without transcription or linkage to external voter or property records, establishing audio-as-data as a practical approach for the study of local politics.
- [313] arXiv:2611.18074 [pdf, html, other]
-
Title: Learning nonlinear interactions and cross-feature couplings with product-unit residual networksSubjects: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
Understanding nonlinear feature interactions is crucial in science and engineering, yet standard multilayer perceptrons (MLPs) often capture such interactions only implicitly, leading to entangled representations that can impair robustness and interpretability. We investigate complex-valued product-unit residual networks (CV-PURe) that integrate multiplicative product units with residual connections to explicitly model cross-feature couplings while stabilizing optimization. We conduct a systematic evaluation on an interaction-driven synthetic benchmark and real-world datasets and evaluate prediction together with input-level interaction recovery under a reproducible protocol. Across five data splits and five training seeds, CV-PURe achieves the lowest mean RMSE on four of six scientific benchmarks, reducing Friedman error by 53.9% relative to MLP. Its Friedman interaction maps assign 60.7% of total strength to the known interacting pair, compared with 31.0% for MLP. On a synthetic task containing third- and fourth-order products and 20 distractors, CV-PURe reduces RMSE from 0.1376 to 0.0797 relative to a complex-valued MLP and recovers all nine true pairwise projections in every training/sampling repetition. Bike Sharing experiments connect improved prediction to the contribution of temporal, calendar, and seasonal context through paired ablations of both PURe variants. Parkinsons Telemonitoring extends the context analysis to repeated measurements. Noise and learning-curve evaluations demonstrate sustained predictive gains under perturbed inputs and limited supervision, while inference measurements quantify computational cost. This article extends our ICCS 2026 paper, Modeling Nonlinear Feature Interactions with Product-Unit Residual Networks (doi:https://doi.org/10.1007/978-3-032-29921-5_6).
- [314] arXiv:2611.18075 [pdf, html, other]
-
Title: Zero-SM NCCL: A Systematic Guide to Designing and Deploying SM-Free GPU CommunicationZhenhao He, Siyuan Shen, Ke Wen, Prajwal Singhania, Shiqing Fan, Youngeun Kwon, Sangkug Lym, Vasudevan Rengasamy, Vishnuswaroop Ramesh, Yicheng Qiang, Craig Belusar, Jeff Hammond, Torsten HoeflerSubjects: Distributed, Parallel, and Cluster Computing (cs.DC)
GPU communication libraries such as NCCL use Streaming Multiprocessors (SMs) for data movement and synchronization, slowing overlapped computation at a cost that grows with interconnect bandwidth. We present Zero-SM NCCL, which communicates within and across scale-up domains without launching any SM kernel. It synchronizes through CUDA stream memory operations, moves data asynchronously with copy engines within a scale-up domain, and drives network RDMA from a stream-ordered CPU proxy. It supports data-movement collectives such as AllGather and AlltoAll, plus one-sided ncclPutSignal and ncclWaitSignal APIs. On GB200 NVL72 and B200 NVL8 systems, Zero-SM collectives reach up to 1.36x higher AllGather bandwidth than NCCL's legacy SM-based kernels and cut the median slowdown of overlapped GEMMs from 33% to 4%, at a higher latency per operation. In Megatron-LM training, they accelerate FSDP by up to 3.5%, and one-sided puts accelerate pipeline parallelism by up to 15%. Zero-SM communication is not a universal win, however. Through an end-to-end study across models and parallelism strategies, we show that it pays off for large transfers in GPU-bound workloads and for deep pipelines with large bubbles, but loses on small messages and host-bound workloads. We distill these findings into guidelines for enabling zero-SM communication in NCCL and building it into other libraries.
- [315] arXiv:2611.18077 [pdf, html, other]
-
Title: Grid-Compliance Testing of Wind Turbines through Virtualized Component Tests: Testing Concept and Test BenchSimon Moritz, Yoga Viknesh Kannan, Florian Andresen, Laura Fuentes Grau, Manuel Fischer, Johannes-Martin Hördler, Sabri Özer, Martin Brennecke, Philipp Kersting, Andreas Thul, Simon Steentjes, Ferdinanda Ponci, Antonello Monti, Rik W. De Doncker, Heike Vallery, Maximilian BaslerComments: 10 pages, 7 figures, 5 tables, peer-reviewed journal articleJournal-ref: atp magazin 08/2026, pp. 48-61, 2026Subjects: Systems and Control (eess.SY)
As the proportion of inverter-connected wind turbines increases, test methods for evaluating their grid behavior are becoming increasingly important. This paper presents a Power Hardware-in-the-Loop (PHiL)-based method for grid connection testing of wind turbines. For the first time, geographically distributed simulations of aerodynamics via the Internet and of the grid model via a fiber-optic connection are integrated into certification-relevant tests. The concept expands upon the methods described in IEC TS 61400-21-4. Initial experimental results confirm high emulation quality and sufficiently low latencies, and meet the requirements for a standards-compliant grid emulation.
- [316] arXiv:2611.18078 [pdf, html, other]
-
Title: TelePhysica: Agentic Reconstruction of Physically Interactive ScenesComments: Ongoing Work. Homepage: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Direct image-to-video diffusion models entangle pixel synthesis with implicit latent motion, precluding persistent 3D mechanical equilibrium, mid-rollout intervention, and counterfactual force control. Conversely, single-pass programmatic 3D reconstruction from an uncalibrated photograph leaves monocular depth and support clearance underconstrained, causing reconstructed scenes to interpenetrate or collapse under gravity. We present TelePhysica, a closed-loop framework that reconstructs an executable, multi-regime physical world from a single photograph to drive stateful 3D interaction and trajectory-grounded video synthesis. A vision-language coding agent synthesizes separable visual meshes and declarative physical models within a two-stage verification loop that pairs pixel-aligned silhouette critique with a zero-force gravitational equilibrium preflight. The verified world compiles into a unified solver coupling rigid bodies, articulated chains, co-dimensional cloth shells, and 3D tetrahedral elastodynamics, whose simulated rollouts condition pretrained video diffusion models through pixel-aligned structural streams. On a 100-scene single-image benchmark, closed-loop execution raises equilibrium-verified world recovery from $0\%$ in single-pass synthesis to $81\%$. Controlled constitutive ablations show over an order-of-magnitude larger localized corner displacement and out-of-plane pleating in finite-element cloth than in matched rigid-plate proxies. Conditioning video diffusion on simulated trajectories increases dynamic motion-mask overlap by $2{\times}\text{--}5{\times}$ and preserves motion direction under counterfactual force reversals where direct image-to-video models fail.
- [317] arXiv:2611.18079 [pdf, html, other]
-
Title: High-order IMEX local discontinuous Galerkin methods with sparse grids for high-dimensional nonlinear parabolic PDEsSubjects: Numerical Analysis (math.NA)
In this work, we develop a high-order IMEX-LDG scheme on sparse grids for solving nonlinear parabolic PDEs in moderate to high dimensions. The method combines the advantages of LDG methods for convection--diffusion problems, IMEX-RK time integrators for stiff systems, and sparse-grid techniques to mitigate the curse of dimensionality. The resulting scheme is fully parallelized in MPI, is robust, efficient, and scalable, making it suitable for a wide range of applications involving high-dimensional nonlinear parabolic PDEs.
- [318] arXiv:2611.18081 [pdf, html, other]
-
Title: Generalized hydraulic neural operators turbocharging sewer digital twinsSubjects: Computational Engineering, Finance, and Science (cs.CE)
Hydraulic digital twins are emerging as the next-generation paradigm for operating and managing sewer systems. Conventional hydraulic digital twinning relies on running physics-based models (e.g., Saint-Venant equations), which are computationally expensive and often too slow for real-time applications. Although surrogates (e.g., neural networks) can be fast, existing approaches are typically non-generalizable, requiring costly retraining for unseen pipes, networks, or flow conditions. Here, we develop a fast, generalized and differentiable multi-branch neural operator trained on solutions of the Saint-Venant equations that predicts in-pipe hydraulics. The learned operator reproduces complex hydraulic dynamics with high fidelity while achieving thousand- to million-fold speedups for online prediction, data assimilation, and diagnostics. Equally importantly, the learned operator generalizes to unseen pipes, networks, and flow conditions without retraining. These advances address key computational barriers to real-time sewer digital twins, opening new opportunities for more responsive flood-risk management and protection of urban water environments.
- [319] arXiv:2611.18084 [pdf, html, other]
-
Title: On Applications of Polynomial-Time Group Isomorphism AlgorithmComments: 10 pagesSubjects: Computational Complexity (cs.CC)
The group isomorphism problem is a fundamental problem in computational group theory and is closely related to many other problems. Recently, Qiao and Zhang proposed a polynomial-time algorithm for finite group isomorphism in the multiplication-table model that computes the coset of all isomorphisms. Based on the algorithm of Qiao and Zhang, we study applications to finite associative ring isomorphism, bilinear map isotopism over finite commutative rings and free three-tensor isomorphism over finite commutative rings. First, we obtain a polynomial-time algorithm that returns the coset of all isomorphisms for finite associative rings. Second, we extend efficient isotopism and isomorphism algorithms for bilinear maps and free three-tensors from fields to general finite commutative rings.
- [320] arXiv:2611.18085 [pdf, html, other]
-
Title: Distributed Output-Feedback Control for Linear Multichannel SystemsSubjects: Systems and Control (eess.SY)
This paper investigates distributed output-feedback stabilization of linear multichannel systems in which agents exchange state estimates without disclosing local control inputs or feedback gains. Remote control actions are treated as unknown inputs and geometrically decoupled from stable quotient-error dynamics. The proposed framework accommodates nontrivial intersections of local loss subspaces by retaining the coupling between the plant and common estimation errors. A Hurwitz condition on the coupled dynamics, together with a sufficient consensus-gain bound, guarantees exponential closed-loop convergence for every fixed feedback selection in a compact family. For polytopic feedback families, a quotient-preserving parameterization of output injections enables joint synthesis of nominal feedback and observer gains via alternating linear matrix inequalities, followed by a convex semidefinite program certifying a common consensus gain. A numerical example demonstrates the effectiveness of the proposed design.
- [321] arXiv:2611.18088 [pdf, html, other]
-
Title: Two-Clustering Regime of Token Dynamics in Causal AttentionSubjects: Artificial Intelligence (cs.AI)
We partially resolve Conjecture 1 of [Karagodin et al., Clustering in Causal Attention Masking (NeurIPS 2024)] and extend it to time-dependent parameter regimes. The conjecture predicts that causal attention dynamics converge to two clusters aligned with the eigenvectors associated with the largest eigenvalue $\lambda_{\max}$ of the value transformation. We prove that this behavior holds when $\lambda_{\max}>0$ and is generally false when $\lambda_{\max}\le0$. The resulting two-cluster regime is common in practice and substantially more challenging than the classical single-cluster setting studied previously. A central contribution of our work is the identification of a rigorous dimension reduction phenomenon: despite evolving in a high-dimensional embedding space, token trajectories collapse onto a two-dimensional structure governed by a small number of dominant directions. Building on this insight, we introduce a new analytical framework--instability within the flow--inspired by instability phenomena in physical systems. This framework enables the analysis of steady-regime behavior in non-monotone, non-autonomous attention dynamics. Our theoretical findings are illustrated through simulations and experimental studies using Group Query Attention (GQA) layers.
- [322] arXiv:2611.18089 [pdf, html, other]
-
Title: Deep Latents, Long Memory: Latent persistence explains temporal correlations in transformer activationsSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
The hidden layers of Large Language Models (LLMs) are believed to encode representations of the data whose abstraction increases with depth. However, in the absence of an operational notion of abstraction, this statement is hard to test quantitatively. Here, we propose an alternative view, building on the principle that the correlations of a sequence reflect the structure of the process that generated it. We define the autocorrelation function of a layer as the average dot product between its activations at two context positions separated by a lag. Latent variables encoded persistently along the context create slowly decaying tails in the autocorrelation, and we find these tails to grow across layers in both synthetic and natural data. In synthetic data with controlled hierarchical latent structure, this growth stems from the encoding of abstract latent variables, which govern long stretches of the input. Indeed, the persistence of latent codes predicts the qualitative structure of autocorrelations and accounts for most of the empirical measures. In real text, we take sparse-autoencoder features as proxies for latent variables, and find that their persistence correlates strongly with the magnitude of autocorrelations across datasets and LLMs. In particular, the persistence time increases with depth, as do autocorrelations, implying that deeper latent features persist for longer context spans. We complement this analysis by explicitly ablating the persistent sparse-autoencoder features and observing a consequent decrease in the autocorrelations.
- [323] arXiv:2611.18090 [pdf, html, other]
-
Title: Bound-preserving finite volume scheme for a coupled parabolic system with nonlinear and nonlocal driftSubjects: Numerical Analysis (math.NA)
We propose a bound preserving finite-volume scheme for a coupled system of partial differential equations with nonlinear, nonlocal drift. The main contribution of this work is twofold: $(1)$ we show that the proposed scheme preserves physically motivated bounds; $(2)$ we prove that the scheme is well-posed and leads to a sequence of approximate solutions converging to the weak solution of the original problem. This setting is motivated by questions related to morphology formation in organic solar cells.
- [324] arXiv:2611.18092 [pdf, html, other]
-
Title: Beyond Scalar Inertia: Frequency-Response-Based Assessment of Finite-Duration Inertial Support in Grid-Forming ConvertersComments: This paper has 6 pages with 5 figures. The target journal will be determined at a later stageSubjects: Systems and Control (eess.SY)
Conventional scalar inertia characterizes only the low-frequency terminal behavior of grid-forming converters (GFMs) and cannot fully predict inertial energy delivery during short-duration frequency disturbances. This letter proposes a frequency-response-based method to assess finite-duration inertial support using complex terminal power--frequency responses, such as those provided by Network Frequency Perturbation (NFP) plots. For GFMs with nonzero steady-state droop, a duration-dependent energy-equivalent inertia is defined after subtracting the ideal static-droop contribution. An exact spectral relationship is established, expressing this metric as a duration-weighted integral of frequency-dependent inertia, without requiring knowledge of the internal controller structure and parameters. The duration dependence of the metric is further shown to encode observable oscillation frequencies and damping ratios. A coefficient-decoupling GFM control with identical scalar inertia and droop across different settings is used for verification.
- [325] arXiv:2611.18094 [pdf, html, other]
-
Title: Controllability-Informed Distributed Electric Propulsion Aircraft Design Using Probabilistic Attainable Moment SetsSubjects: Systems and Control (eess.SY); Optimization and Control (math.OC)
Integrated aircraft design combines models of aerodynamics, propulsion, mass, and mission performance in one design loop, so that the disciplines can be traded against each other from the earliest stage. For configurations with distributed electric propulsion (DEP), part of the control authority comes from the propulsion system itself, which directly affects the required size of the tail and should therefore be known while the aircraft is being sized. This paper presents a controllability measure that is evaluated for every design in the loop and returned to it as a design signal. The attainable moment set of each candidate design is computed from the surrugate models of the loop and it is compared with a required moment set. Since the low-fidelity surrugate models are uncertain, the comparison is formulated with probabilistic attainable moment sets: the uncertainty is propagated to the boundary of the set, which yields a margin for every direction in the moment space and the probability of violating the requirement. The design loop thereby informed whether a configuration provides sufficient control authority, on which axes it falls short, and how this changes with the design variables. The approach is demonstrated on a DEP regional aircraft, where the signal is used to assess the size of the vertical tail.
- [326] arXiv:2611.18096 [pdf, html, other]
-
Title: VEPO: Mitigating Multilingual Hallucinations in LVLMs via Visual Evidence-guided Preference OptimizationSubjects: Computation and Language (cs.CL)
Large vision-language models (LVLMs) achieve strong multimodal performance, yet they remain prone to object hallucination---generating answers that are linguistically fluent but visually ungrounded. This issue is especially severe in low-resource languages, where weaker cross-modal alignment causes models to rely more on language priors than on image evidence. Existing remedies either require costly multilingual annotations or external translation/reward models, or intervene at inference time in ways that can distort the output language. We propose VEPO(Visual Evidence-guided Preference Optimization), a training-time method that transfers reliable visual grounding from English to other languages without external models. VEPO first localizes key visual evidence using English queries via mid-layer cross-modal attention, then constructs preference pairs that share the same correct answer but differ in evidence sufficiency (full image vs.\ key-region masked image). Direct preference optimization on these pairs encourages low-resource inputs to prefer answers supported by complete visual evidence. Experiments on multilingual POPE and MME with Chinese, French, and Hindi show consistent hallucination reduction and perception gains, with the largest improvements on the lowest-resource language. Analyses further confirm that VEPO sharpens question-relevant attention and that evidence-guided masking outperforms random masking.
- [327] arXiv:2611.18097 [pdf, html, other]
-
Title: AG-JEPA: Anatomically Guided Joint Embedding Predictive Architecture for Neurodegenerative Representation LearningAzad Singh, Ayantika Das, Amog Rao, Aiman Farooq, Anirban Porya, Ananya Shukla, Sowbhagya Suresh, Srinivasan Iyengar, Mohit Jain, Mercy Prasanna RanjitSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Brain MRI can reveal early signs of Alzheimers disease (AD) before measurable cognitive decline becomes apparent. However, during the transition from cognitively normal (CN) to mild cognitive impairment (MCI), the disease-related signals are often subtle and confined to only a few brain regions. Existing Joint Embedding Predictive Architecture (JEPA) models learn general structural representations but do not explicitly capture the anatomical regions affected in AD. We propose Anatomically Guided JEPA (AG-JEPA), which introduces anatomical tokens that interact with patch tokens during pretraining to learn region-specific representations from brain regions relevant to neurodegeneration. Building on this representation space, we further propose Prog-AG-JEPA, which models disease progression trajectory as a conditioned latent displacement within this space. We evaluate AG-JEPA across five downstream tasks spanning Alzheimer's disease diagnosis and progression prediction. AG-JEPA improves diagnosis and MCI prognosis, achieving an 8.4 percentage point AUROC gain over the strongest baseline on CN/MCI classification and a 7.1 percentage point gain on out-of-distribution CN/AD classification on the MR-RATE dataset, highlighting its effectiveness across both early-stage detection and cross-cohort generalization. For disease progression modeling, Prog-AG-JEPA consistently outperforms all baselines on future diagnosis and brain volume change prediction, demonstrating the value of anatomically grounded neurodegeneration representations for diagnosis and prognosis.
- [328] arXiv:2611.18098 [pdf, other]
-
Title: Do Machines Learn Deviance From Their Peers? A Randomized Test of Differential Association With AI AgentsSubjects: Artificial Intelligence (cs.AI)
Differential association theory and its extension, social learning theory, hold that rule-breaking, and the definitions that make it acceptable, are learned through interaction with others. Because people choose their friends, this claim has been hard to test. We addressed this selection problem with AI agents, which learn only through communication and can be assigned to peers at random. In a preregistered experiment, 420 teams of three agents (1,260 agents; seven AI models) worked on coding tasks that could not be solved honestly. A scripted teammate said that it had cheated (deviant peer), that it had reported the problem honestly (honest peer), or said nothing (silent peer). Half of the teams were told that an auditor read their chat. Compared with a silent peer, a deviant peer increased cheating (66.2% vs. 62.9%; OR = 1.96) and an honest peer reduced it (55.7%; OR = 0.35). Among agents that cheated, a deviant peer increased claims that cheating was normal (13.7% vs. 3.4%), often as the team's norm. Monitoring did not suppress prohibited discussion; it doubled it. Peers mattered only for models that had not yet decided. AI agents offer a new way to test criminological learning theories without selection effects.
- [329] arXiv:2611.18099 [pdf, html, other]
-
Title: EcoVGGT: Pinned-Key Inference for Efficient and Streamable Feed-Forward 3D Reconstruction from Long VideosSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Feed-forward multi-view transformers such as VGGT recover the cameras and geometry of a video in one pass, but the cost of their cross-frame attention grows quadratically with the frame count, and their accuracy drops on long videos. Most remedies approximate attention over all frames within a budget or select a few keys by their content. We hypothesize that the keys of a long video can be chosen without looking at its content, and our experiments on VGGT and $\pi^3$ support it: a few dozen evenly spread frames can serve as the keys of every frame. We present EcoVGGT, a training-free inference framework whose keys form a Pinned Key Set of 31 such frames. A pin pass stores their tokens in a Scene Cache, and the other frames read it in independent batches that reproduce joint inference with these keys exactly, at a constant cost per frame. An Online Stream updates the pinned frames during recording. Six middle global blocks align the frames and form a reusable VGGT module: the Alignment Score reads their attention to flag failures, and Alignment Refocusing reduces them by fine-tuning only these blocks (3.6 GPU hours). EcoVGGT ($\pi^3$) matches the success rate of $\pi^3$ on 1,000-frame ScanNet videos at 13.8 times its speed. With refocusing, EcoVGGT (VGGT) raises the success rate of VGGT from 70% to 82% at 8.5 times its speed on 500 frames. EcoVGGT reconstructs 4,000 frames on one GPU, and EcoVGGT ($\pi^3$) streams 1,000 frames at 18.6 frames per second with the lowest online absolute trajectory error.
- [330] arXiv:2611.18101 [pdf, html, other]
-
Title: KnowVis: Dual-View Evaluation and Diagnosis of World-Knowledge Grounding in Text-to-Image ModelsComments: Accepted to NeurIPS 2026 E&D Track. Codebase: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV)
Text-to-image (T2I) models have made substantial progress in visual realism, aesthetic quality, and instruction following. However, real-world prompts often go beyond explicit visual descriptions and require implicit facts, structured knowledge, and domain-specific commonsense. Existing evaluations mainly focus on explicit prompt-to-image semantic alignment or specific knowledge subdomains, lacking unified evaluation and diagnosis of world-knowledge grounding in T2I models. We introduce KnowVis for dual-view evaluation and diagnosis of world-knowledge grounding in T2I generation. Built on a hierarchical knowledge taxonomy with case-level annotations, KnowVis combines two complementary benchmark views with a structured MLLM-based VQA judge protocol. The Skill-Tree Benchmark supports fine-grained atomic knowledge evaluation, while the Generalized Benchmark examines open-ended knowledge composition and expression. The protocol uses Required VQAs to assess atomic knowledge correctness in the Skill-Tree Benchmark, and combines Required, Optional, and dynamically discovered knowledge expressions with human-judge collaborative evaluation in the Generalized Benchmark. We validate the protocol's reliability on the Skill-Tree validation split, enabling scalable evaluation and diagnostic error attribution. Experiments show that KnowVis reveals notable gaps in world-knowledge grounding across existing T2I models and provides structured diagnostic signals for knowledge-intensive image generation.
- [331] arXiv:2611.18104 [pdf, html, other]
-
Title: Contagious Fingerprints: When LLMs Inherit Another Model's IdentitySubjects: Computation and Language (cs.CL)
The fingerprint by which a classifier recognizes the LLM behind a text is usually treated as a fixed property of the model. We show that it is contagious: a model that continues a conversation begun by another model inherits that model's fingerprint. This undermines audits of API relays, which can serve the first turns of a conversation with the model a customer pays for and silently hand the rest to a cheaper one. After such a handoff, detection of the substitute by a text classifier drops from 0.97 to 0.72 (DeepSeek replacing Claude Sonnet) and from 0.87 to 0.54 for a same-vendor downgrade (Haiku), and retraining on handoff data repairs only substitutes seen in training. Preregistered experiments trace the contagion to imitation of the history's style. History also moves a model away from its own fingerprint: active probes are not camouflaged, but when calibrated on fresh conversations they flag \emph{honest} models on long contexts at rates up to 1.00, and a standard MMD equality test rejects all nine models we test. Both failures come from comparing a history-dependent reply with a history-free reference. We propose \textsc{Lineup}, which samples references from the claimed model \emph{on the same history} and tests the audited reply by its rank among them. \textsc{Lineup} controls false positives by exchangeability and needs no fitted threshold or log-probabilities. With contemporaneous references on two closed models, its false-positive rate is 0.056 (preregistered bound 0.10), and it detects 0.81 of substitute sessions across six pairs after four audited turns.
- [332] arXiv:2611.18110 [pdf, html, other]
-
Title: PoliMop: Policy-in-the-Loop Motion Planning in 3D Spaces for Mobile RobotsComments: 8 pages, 5 figures. Videos and code: this https URLSubjects: Robotics (cs.RO)
Reaching a goal state in a complex environment often requires a robot to overcome challenging terrain and obstacles. Which routes are accessible depends on the robot's physical design, the control policy it uses, and how it modifies its surroundings. We present a kinodynamic rapidly-exploring random tree (RRT) planner that uses physics simulation to exploit skills encoded in the control policy, which allows it to find a path in the scene with respect to the robot's capabilities. The planner samples states from a reduced state space in 3D around the scene and expands the RRT toward each sample using policy rollouts. We guide sampling with a preplanned path for a simple mass-point model to focus exploration on promising regions leading toward the goal state, reducing planning time by up to 63%. Simulation experiments with quadrupeds, humanoids, and wheeled robots demonstrate that our approach supports drop-in integration of various robots and control policies, exploiting their physical and policy capabilities without modifying the underlying planning algorithm. Predefined environmental modifications, such as placing a plank across a gap to build a bridge, will automatically be picked up by the planner as the robot's reachable space expands.
- [333] arXiv:2611.18111 [pdf, html, other]
-
Title: Multi-GPU Tridiagonal Partition AlgorithmComments: 3 pages, 3 tables, Mathematical Modelling and Computational Physics 2026 international conference (MMCP'2026), Dubna, RussiaSubjects: Distributed, Parallel, and Cluster Computing (cs.DC)
This paper presents a multi-GPU version and a multi-GPU recursive version of the parallel partition method for systems of linear algebraic equations with tridiagonal coefficient matrices and their implementations in CUDA. Computational experiments on one and two GPU cards, for different system sizes and with different number of recursive steps are conducted. The results are analyzed and discussed.
- [334] arXiv:2611.18112 [pdf, html, other]
-
Title: Any Channel Count, One Token Stream: Incorporating Spatial Audio Understanding and Reasoning into Large Audio-Language ModelsComments: 27 pages, 5 figures. Code: this https URL. Dataset: this https URL. Paper experiments use the 20,492-question SpheraQA version; the public release contains 22,197 questionsSubjects: Sound (cs.SD)
Audio phase relations encode source locations, yet existing large audio-language models (LALM) reduce audio to single-channel content to stabilise general audio understanding, leaving spatial information for direction, distance, and source-relation questions unused. We present SpheraAudio, which accepts recordings with any channel count -- mono, stereo, First-Order Ambisonics (FOA), or higher -- and maps them into a single token stream. A spatial encoder turns complex multichannel spectra into spatial tokens interleaved with audio tokens at corresponding temporal positions. Spatial Multimodal RoPE (SMRoPE) makes each token's modality identity and temporal slot visible to rotary attention without altering the language model's original RoPE. Training runs in three stages -- spatial-encoder pretraining, projector alignment, and joint supervised-and-reinforcement learning -- progressing from geometric supervision to language-based spatial decisions, with answers trained as explicit spatial chain-of-thought: step-by-step geometric reasoning over the heard scene. Current benchmarks tie questions to a fixed channel-count format and score answers without spatial reasoning evidence. We introduce SpheraQA, which spans mono, stereo, and FOA recordings and organises questions hierarchically from single-attribute perception through cross-source relations to multi-step reasoning, scored programmatically against simulated scene geometry and real-recording annotations with step-by-step solution references. Across SpheraQA and public spatial-audio benchmarks, SpheraAudio sets a new performance standard for spatial audio understanding. It supports high-precision localisation, tracking, and reasoning over static and moving sound sources while preserving strong general-audio capabilities. We will release code, SpheraQA, and evaluation metrics to support reproducibility.
- [335] arXiv:2611.18113 [pdf, html, other]
-
Title: Folding Language Models into ThemselvesComments: 14 pages, 4 figures, 9 tables. Code: this https URLSubjects: Computation and Language (cs.CL)
Knowledge distillation usually trains a compact student from scratch, which needs a large amount of data even though the teacher's weights already encode most of what the student has to learn. In this work, we explore building the student directly from the teacher's weights. We propose Concertina, which folds every group of N adjacent teacher layers into one shared block that is executed N times. Each loop step carries a low-rank correction initialized from the layer it replaces, and a per-token gate mixes these corrections, to let every token recover the behavior of the original layers from a single shared block. We further design a distillation data recipe that balances general, mathematical, academic and targeted multiple-choice data, to recover all capabilities of the teacher together rather than one at the expense of another. On Qwen3-4B, a 2.27B Concertina model distilled on 22B tokens retains about 90% of the teacher on commonsense and mathematical reasoning and 84% on average across eight benchmarks. It outperforms same-size dense and layer-sharing baselines, as well as Qwen3-1.7B, which is trained on over a thousand times more data. These results suggest that a compact model need not be retrained at scale, since folding the teacher into itself and distilling it on the right data recovers most of its capabilities. Code is available at this https URL.
- [336] arXiv:2611.18114 [pdf, html, other]
-
Title: Interpretable weight-space programming builds transparent transformers and brain encoding modelsYash Mehta, Chandan Singh, Weijia Xu, Richard J. Antonello, Adeen Flinker, Nima Mesgarani, Michael F. Bonner, Nebojsa Jojic, Jianfeng GaoSubjects: Computational Engineering, Finance, and Science (cs.CE); Artificial Intelligence (cs.AI)
Neural networks are powerful predictors, yet their opacity limits their usefulness in science, where the goal is explanation rather than mere prediction. In neuroscience, for example, deep encoding models accurately predict brain activity in response to natural stimuli, but struggle to reveal which stimulus features drive those predictions. Here we introduce interpretable weight-space programming (iWSP), a method in which an LLM agent iteratively builds interpretable models by explicitly programming their weights and features in code, without gradient-based training (backpropagation). Without backpropagation to tune parameters, the agent must draw on its prior knowledge and reason over previous trials to hypothesize, hand-craft, and refine human-interpretable computations mirroring the workflow of a human scientist. LLM audits confirm the interpretability of the hand-crafted models. We first validate iWSP by tasking it with constructing transformers for canonical benchmarks such as arithmetic and sentiment classification, finding that LLM agents can successfully craft compact, accurate networks directly. Next, we apply iWSP to build neural encoding models across vision, language, and audition. In the visual cortex, iWSP outpredicts existing interpretable baselines, uncovers distinct computations across early/higher visual areas, and generalizes to unseen data. In language fMRI, iWSP surfaces a confound in the data and in auditory ECoG, it systematically discovers interpretable acoustic features that predict neural activity. Together, these results establish iWSP as a general paradigm for aligning high predictive performance with human-interpretable scientific discovery.
- [337] arXiv:2611.18119 [pdf, html, other]
-
Title: Query Processing with eBPF-Offload on Computational StorageAlexander Krause, Niclas Hedam, Johannes Pietrzyk, Dirk Habich, Wolfgang Lehner, Sangjin Lee, Alberto Lerner, Philippe BonnetSubjects: Databases (cs.DB)
The idea of pushing computation to storage devices has been explored for decades. In recent years, research has focused on computational storage, i.e., devices that allow a host to execute functions as I/O commands on local data. Today, the NVMe and SNIA computational storage standards share an execution model in which the host invokes functions on the device, on data that the host has moved into device memory. Both standards single out eBPF as a format for functions offloaded by the host. In this paper, we study whether this form of computational storage is a good match for query processing. We present the design and implementation of Delilah, a prototype based on an open-hardware device that implements the standard execution model on an actual device and runs eBPF programs calling device-resident functions. As an initial step, we hand-coded a portion of SSB Q3.3 for execution on the device, aiming to integrate computational storage with a database system's storage manager. The series of experiments that we ran showed that offloading gains little over host execution, execution degenerates into column-at-a-time processing, and time is spent staging and moving data in device memory. Because these roadblocks stem from the execution model rather than from the query, we did not pursue integration further. We detail these limitations and lay the groundwork for a different computational storage model better aligned with data-intensive workloads.
- [338] arXiv:2611.18121 [pdf, html, other]
-
Title: Optimizing Drilling Trajectories in Multi-Material Structures Using Quantum AnnealersComments: 7 pages, 2 figures, 3 tables. Paper submitted to IEEE QCNC 2027 conferenceSubjects: Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET)
In this work we present a quantum-hybrid optimization approach for planning drilling trajectories on multi-material structures, with the objective of minimizing total energy consumption. The proposed formulation models the drilling process as a constrained routing problem in which each hole must be visited while accounting for heterogeneous material properties, material-dependent drilling costs and displacement energy, as well as tool-wear penalties. Our method incorporates realistic industrial constraints, including open-path routing and mandatory grouping of holes belonging to the same material to avoid frequent tool-bit changes. The modeled problem is solved using D-Wave's Leap Constrained Quadratic Model Hybrid Solver. To evaluate performance, we benchmark instances ranging from 10 to 55 holes with varying material distributions and cost matrices. The results highlight the capability of quantum-hybrid optimization to handle complex multi-physics constraints and to generate energy-efficient drilling sequences under realistic manufacturing conditions.
- [339] arXiv:2611.18123 [pdf, html, other]
-
Title: Evaluating Language Models as Spatial Priors for Object-Goal NavigationSubjects: Robotics (cs.RO)
Semantic exploration methods for object-goal navigation use large language models (LLMs) to score how strongly a target object is associated with a room type or with a nearby object. This paper considers obtaining from the LLM a spatial prior instead: a distribution over the position of the target in a room. We evaluate whether two open-weight LLMs, Qwen3.5-27B and Gemma-4-31B-it, can provide such a prior. Using real room layouts from CubiCasa5K and ScanNet, we measure the likelihood the models assign to real and perturbed layouts, and the layouts they generate. The object frequencies and room sizes of the generated rooms follow the room type, but the proximity of the objects to the walls does not follow the differences between room types found in real rooms. The likelihood distinguishes real layouts from perturbed ones partly through their geometry and partly through the agreement between the coordinates and a field that states the wall each object touches; for the movable furniture of ScanNet, it is higher for layouts with all objects piled near the room center than for the real layouts in most rooms. When an object is given in the prompt, the models place a related object relative to it. When the wall of each object is given, they place most objects against that wall, but do not keep all objects inside a given room rectangle. As a prior for object search, the generated placements reduce the searched area relative to a uniform prior, but by less than a third of the reduction obtained with a prior built from real layouts, and by less than two thirds when both priors are spatially smoothed. These results support the use of LLMs for object-room associations and for writing coordinates under constraints determined by another component, followed by a geometric check, but not as a source of placement priors where real layouts are available.
- [340] arXiv:2611.18125 [pdf, html, other]
-
Title: Efficient Poincaré Convolutions through Patch-Statistics FactorizationComments: Includes supplementary material. Code: this https URLSubjects: Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
Published implementations of the Poincaré convolution in Hyperbolic Neural Networks++ (HNN++) map every overlapping image patch to the Poincaré ball before evaluating its score. We show that the clipped score depends on each patch only through two statistics, its squared norm and its inner products with the filters, which a native convolution and a single-channel box filter compute directly. The factorized layer therefore evaluates the specified score exactly, without constructing ball-valued patches, and its single-precision outputs and gradients agree with double-precision evaluations at least as closely as those of the explicit reference. On an RTX A6000 under two versions of JAX, complete training updates of a Poincaré U-Net and a Poincaré ResNet-20 are 1.5 to 1.8 times faster than the faster of two explicit implementations that map patches to the ball (1.2 to 1.4 times against an explicit implementation that skips this mapping); on a V100, the gain over the gather-based one persists as the ResNet-20 is widened eightfold. In 15 paired training runs per task, updates are 2.0 and 1.8 times faster than a reference that follows the published execution order, with no consistent change in final segmentation or classification quality, and peak memory falls by 38 to 55% in a U-Net scaling study. Replacing the convolutions of the official PyTorch Hyperbolic U-Net makes its training step 1.4 times faster with half the memory. Evaluating an existing layer from its patch statistics thus makes Poincaré convolutional networks cheaper to train in the tested models and software stacks.
- [341] arXiv:2611.18126 [pdf, html, other]
-
Title: Hello Again: Exploiting Replay Protected Monotonic Counters to Brute-Force Windows Hello PINsChristian Werling, Matti Finder, Hans Niklas Jacob, Julian Prein, Niclas Kühnapfel, Jean-Pierre SeifertComments: This is a slightly updated version of the paper published at ACM CCS '26 (this https URL). The section Changes from the CCS '26 Version lists all changesSubjects: Cryptography and Security (cs.CR)
Windows Hello is the primary authentication mechanism on over one billion Windows 11 devices, supporting PINs as short as four digits. Such low-entropy PINs are viable only because the TPM's anti-hammering protection mitigates brute-force attacks. Most modern x86 platforms rely on firmware TPMs (fTPMs) that persist their non-volatile (NV) state, including the lockout state, to SPI flash. On newer platforms, rollback prevention of this state is delegated to the flash chip via Replay Protected Monotonic Counters (RPMCs), a JEDEC-standardized mechanism operating over a physically accessible bus.
We present the first practical security analysis of RPMC in shipping x86 systems. Using low-cost interposer hardware, we identify multiple implementation flaws, two of which fundamentally undermine RPMC's security model: First, Winbond SPI flash - the leading NOR flash vendor with over 25% market share - leaks its HMAC key under voltage glitching, enabling arbitrary counter forgery. Second, AMD's NV controller does not verify RPMC increments, so an interposer can suppress them unnoticed. Moreover, AMD platforms without RPMC flash appear to lack NV rollback protection altogether: on two AMD desktops, reflashing an old image resets the fTPM's lockout counters.
Both flaws enable rollbacks of the fTPM's NV state (AMD, Intel, or Pluton), including its lockout counters. We demonstrate this end-to-end on locked, encrypted hardware: our setup finds a 4-digit Windows Hello PIN in 40 hours instead of 69 days. A compromised PIN also grants access to device-bound keys and Hello-based enterprise SSO. The same primitive defeats BitLocker's TPM+PIN pre-boot protector, reducing the worst-case time for a default-length 6-digit PIN from 19 years to 20 days. - [342] arXiv:2611.18127 [pdf, html, other]
-
Title: PRISM: A Post-Prediction Reliability Framework for Evidence-Guided Correction and Selective AbstentionComments: 7 pagesSubjects: Networking and Internet Architecture (cs.NI); Machine Learning (cs.LG)
A model may produce a prediction with high confidence and still get the outcome wrong. This is a challenge in sequential applications, where the system evolves over time and deployment data may differ from the data used during training. To address this issue, we propose PRISM (Prediction Reliability and Intelligent Self-Correction Mechanism), a framework that evaluates prediction reliability after the initial output is generated. It estimates the risk of error and examines predictions identified as potentially unreliable. For these cases, PRISM analyses available signals, retrieves similar situations from historical data, and verifies whether an alternative prediction is supported by the evidence. It can then retain the original prediction, apply a verified correction, or abstain when the evidence is insufficient.
We evaluate PRISM on second-innings match prediction in the Indian Premier League (IPL). Ball-by-ball records are converted into over-level observations to capture changes in match conditions throughout an innings. To preserve temporal order, earlier IPL seasons are used for training, IPL 2024 for validation, and IPL 2025-2026 for out-of-time testing. The dataset contains 8,665 over-level observations from 475 matches, with the test partition comprising 2,568 observations from 142 matches. In one experimental record, Logistic Regression, Random Forest, and HistGradientBoosting achieve accuracies of 65.2%, 68.4%, and 70.2%, respectively. The corresponding PRISM evaluation reports a selective accuracy of 82.4% at 88.5% coverage, with a Brier score of 0.126 and an expected calibration error (ECE) of 0.019. - [343] arXiv:2611.18130 [pdf, html, other]
-
Title: JFinBench: A Continuously Refreshable Japanese Financial Benchmark from the Latest Certification ExamsComments: 32 pages, 14 figuresSubjects: Artificial Intelligence (cs.AI)
Large language models (LLMs) are increasingly applied to finance, creating a need for benchmarks that evaluate their capabilities across financial systems with different regulations and practices. However, existing financial benchmarks provide limited coverage of country-specific financial systems, and they are difficult to keep up to date as released questions leak into training data. We introduce JFinBench, a Japanese financial benchmark comprising 321 examples from the 2026 editions of six regularly administered certification exams covering accounting, tax, financial planning, and banking. To construct it, we develop a human-in-the-loop agentic pipeline in which a coding agent writes exam-specific conversion code and a review page, while a human verifies every example and feeds corrections back to the agent. This pipeline adapts to heterogeneous exam formats and reduces the effort of incorporating new exam editions, enabling periodic refreshes that mitigate data contamination. Evaluating 12 recent LLMs, we find that Claude Fable 5 and GPT-6 Astra exceed 93% overall, while even the best open-weight model, Kimi K3.1 Pro, trails Claude Fable 5 by 12 percentage points. When models can additionally use a terminal and web access, they retrieve the knowledge they lack and every model improves. The gap narrows but about 4 points remain, and the ranking of models stays largely the same, suggesting that knowledge held in the model itself remains valuable even with tools. We release the dataset, the evaluation code, and the construction pipeline.
- [344] arXiv:2611.18132 [pdf, html, other]
-
Title: Parameterized Weakest-Precondition Calculus for Fine-grained Properties of Quantum LoopsComments: 56 pages, 4 figuresSubjects: Logic in Computer Science (cs.LO); Programming Languages (cs.PL)
Quantum loops may execute arbitrarily many iterations, and understanding their runtime behavior requires knowing how often they terminate, how long terminating runs take, and how common very long runs are. We develop a compositional weakest-precondition framework for analyzing these properties on possibly infinite-dimensional state spaces. It uses parameterized predicates: families of quantum effects indexed by a positive observation scale. These families retain runtime information even when the expected runtime is infinite and can be reused through sequential composition and nested loops. We prove that weakest-precondition equations for loops have unique solutions in the parameterized predicate domain. From the asymptotic behavior of the derived weakest preconditions, we then recover these runtime properties and quantify the influence of very long terminating executions. For the classic discrete-time quantum walk on the one-dimensional infinite line with one absorbing boundary at the origin and the Hadamard gate as its coin operator, we verify its termination probability $2/\pi$, conditional mean $\pi/2$, and logarithmic growth of the conditional capped second moment. We then reuse the walk's specification to verify the termination probability and conditional mean runtime of an outer loop that repeatedly invokes it.
- [345] arXiv:2611.18133 [pdf, html, other]
-
Title: Forget the Lineage: Forward-Secure, Lineage-Complete Cryptographic Erasure for LLM-Agent MemoryComments: 14 pages, 2 figures. Submitted to the Journal of Information Security and Applications. Code and LLM-call logs: this https URL (archived at this https URL)Subjects: Cryptography and Security (cs.CR)
Large language model (LLM) agents keep a long-term memory that the model itself writes: summaries, extracted facts and embeddings derived from what users said. Deleting a message is therefore not enough: its derivatives, storage residue and backups survive. We define lineage-complete forward-secure erasure: after a forget request, an adversary that later obtains all storage and the agent's current keys learns nothing about the forgotten input or any memory derived from it. We present LinEra, which encrypts each memory under a key derived from its parents' keys along the derivation graph; source keys come from a puncturable pseudorandom function, so forgetting is a single puncture and every descendant becomes unreadable without being located. A sound lineage rule records as parents everything the writer saw, which confines all influence of a source to its descendants. We prove security and lineage integrity and integrate LinEra with mem0. With real LLMs from two families (1.5B to 7B parameters) writing the memory of LoCoMo conversations, an attacker holding every file and the current keys recovers, on all ten conversations written by the largest model, 70 of 71 forgotten secrets after logical deletion, 67 with per-object or epoch keys, and none with LinEra; the same holds for every writer tested. The price is a larger forgetting radius, which follows a simple law in the writer's fan-in: forgetting 1% of turns erases 73.0% of derived memories with session-level summaries but only 3.8% with four-turn windows, which also answer better; re-deriving restores what is lost.
- [346] arXiv:2611.18137 [pdf, html, other]
-
Title: Rate Optimal Streaming Codes Under an Extended Delay Profile for Three-Node Relay NetworksComments: 27 pagesSubjects: Information Theory (cs.IT)
We study streaming codes for a three-node relay network with burst lengths $b_1,b_2$ and an end-to-end decoding delay $T$. Let $u=\max\{b_1,b_2\}$ and $v=\min\{b_1,b_2\}$, and write $T-v=pu+q$, where $T\ge u+v$ and $0\le q<u$. We develop two binary constructions, SR and SR-2, using extended delay profiles that allow the relay to forward causally invertible combinations of current and past source symbols. The SR construction attains the non-adaptive rate upper bound whenever \[ \frac{T-u-v}{2u-v}\le\left\lfloor\frac{T-u-v}{u}\right\rfloor. \] SR-2 enlarges this range under the sufficient condition \[ q(v-q)\le(p-1)u(u-v). \] These constructions complement the known divisibility and large-delay regimes. We also establish two limitations of systematic ordinary coordinate forwarding. First, it cannot attain the rate upper bound when $0<T-u-v<v$. Second, for equal burst lengths $b$, it attains the bound if and only if $b\mid(T-2b)$; a quantitative upper bound gives a strictly positive rate gap in every nondivisible case. These converses allow nonlinear components and predetermined time variation within the stated ordinary-forwarding architecture. Existing adaptive schemes for equal bursts approach the same rate upper bound asymptotically using time-varying convolutional codes and lie outside the non-adaptive setting considered here.
- [347] arXiv:2611.18139 [pdf, html, other]
-
Title: PUMA-Former: Uncertainty-Aware Agentic Computation Offloading in MEO-LEO Satellite Networks With Incomplete InformationSubjects: Networking and Internet Architecture (cs.NI); Artificial Intelligence (cs.AI)
Satellite edge computing distributes mobile workloads across moving processors whose service capacity must be coordinated over time. In medium Earth orbit (MEO)--low Earth orbit (LEO) networks, periodic reports leave CPU states stale or missing, making offloading sensitive to uncertain capacity and shared-destination congestion. We propose PUMA-Former, an uncertainty-aware agentic framework for cooperative offloading under incomplete information that balances completed work, service delay, and execution cost. A MEO coordinator maintains distinct-report memory and closes the loop between available evidence, capacity estimates, allocation commands, and execution feedback. Latent-delay attention weights plausible measurement-age hypotheses, and historical innovations refine CPU estimates around the latest received report. Queue-error-aware fusion incorporates auxiliary pressure while accounting for backlog uncertainty. The inferred resource moments condition a structured Transformer allocator for joint destination, partition, bandwidth, CPU, and power decisions. Bounded risk-aware calibration then adjusts workload splits to reflect uncertainty and shared destination load. In traffic-driven development evaluation, PUMA-Former increases admitted compute-work completion by 8.62%, reduces mean modeled service delay by 34.78%, and reduces mean episode-P95 delay by 71.08% relative to ATLA-adapt. It also improves completion and normalized energy over fixed-half round-robin offloading. Service gains over the learning baselines involve higher normalized energy.
- [348] arXiv:2611.18140 [pdf, html, other]
-
Title: Large Language Model-Empowered Intelligent Agent for Adaptive Fast-Charging Control of Lithium-Ion BatteriesComments: 9 pages, 5 figures, 6 tablesSubjects: Artificial Intelligence (cs.AI)
Fast charging of lithium-ion batteries is a multi-objective control problem that requires reducing charging time and sustaining high energy throughput while suppressing accelerated degradation and safety-critical electrochemical reactions. Conventional protocols are easy to implement but have limited ability to adapt to variations in battery chemistry, operating conditions, and aging state. Model-based optimization and conventional data-driven control often require accurate models, costly iterative search, or task-specific policies. We develop a large language model (LLM)-empowered intelligent agent for adaptive battery fast-charging control. In this closed-loop framework, the agent interprets structured battery states, generates executable charging actions, observes the resulting battery response, and continuously adjusts subsequent decisions. We construct a large-scale simulation corpus with PyBaMM across five battery chemistries and 11 parameter sets and convert the resulting state-action-response trajectories into structured process data for agent training. Our two-stage strategy combines supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO). SFT establishes battery-control semantics and feasible decision priors from simulation-derived demonstrations. GRPO aligns the agent with a multi-objective control objective covering energy throughput, degradation mitigation, and operational constraints. Closed-loop simulations show that, under matched energy-throughput requirements, the agent achieves longer cycling life than conventional charging baselines and optimization-based strategies across heterogeneous battery configurations. This study provides a simulation-grounded proof of concept for using LLM agents as sequential decision makers in battery management and takes a step toward language-model-enabled control of physical energy systems.
- [349] arXiv:2611.18142 [pdf, html, other]
-
Title: When Copying Is Hard to Beat: A Negative Result in Clinical Assessment ForecastingComments: Accepted at FMTS (Foundation Models for Temporal Systems), a workshop at NeurIPS 2026, SydneySubjects: Machine Learning (cs.LG)
We report a negative result in next-day clinical assessment forecasting: the evaluated models show no consistent advantage over strong persistence-based baselines. Clinicians can view the previous day's questionnaire answers, a plausible source of documentation persistence. Predicting 45 items on 29,715 ICU day-pairs with patient-disjoint splits, a physiology-informed gated model improves item-macro-F1 over a one-line clinical rule by +0.0019 [+0.0006,+0.0031], while losing accuracy. No macro-F1 advantage is established when the physiology window ends at the previous form's reference time (+0.0001 [-0.0006,+0.0008]), or when scoring excludes the 68.5% of cells derived from parent answers; combining both changes yields a small disadvantage. Four checkpoints across three families underperform persistence, and a fine-tuned one differs from it on only two cells. A categorical read-out raises changed-cell accuracy eightfold, but a state-only read-out scores higher overall than forecast-only decoding and matches a transition matrix. This negative result highlights how the documentation process, input cutoff, observation mask and decoding rule shape conclusions about forecasting recorded assessments.
- [350] arXiv:2611.18143 [pdf, html, other]
-
Title: DRA-HTR: Disentangling Representation and Adaptation for Handwritten Text RecognitionSubjects: Computer Vision and Pattern Recognition (cs.CV)
Handwritten text recognition (HTR), particularly when applied to historical documents, remains challenging due to the scarcity of annotated data and the substantial domain shifts caused by variations in handwriting styles, document degradation, acquisition conditions, and linguistic characteristics. While large amounts of unlabeled target-domain data are often available, effectively exploiting them requires adaptation strategies tailored to the specific requirements of HTR. We propose DRA-HTR, an unsupervised generative domain adaptation pipeline specifically designed for HTR and built upon the DRANet framework. The proposed approach combines source-domain pretraining, unsupervised adaptation to the visual characteristics of the target domain, relying on the generation of target-style source samples for task-specific training. We further investigate the integration of post-HTR correction to complement visual adaptation with language adaptation, relying on large language models, which we introduce as DRAL-HTR. The proposed framework is evaluated across seven HTR datasets spanning four languages, using a single set of hyperparameters for all source-target combinations. The recognition model achieves state-of-the-art performance on four datasets and competitive results on the remaining three. In addition, the proposed DRA-HTR establishes a new state-of-the-art for domain adaptation across all seven datasets and substantially improves overall robustness and training stability compared to the original DRANet framework: it reduces the Character Error Rate (CER) in average by 13.52 percentage points, and the average standard deviation decreases from 7.23 to 1.11 percentage points. Finally, post-HTR correction further reduces the CER by 2.16 percentage points.
- [351] arXiv:2611.18144 [pdf, html, other]
-
Title: STAMP: Manufacturing-Informed Generative Design and Shape Optimization of Sheet PanelsComments: 25 pages, 11 figures, 7 tables, including supplementary materialSubjects: Machine Learning (cs.LG)
Generative design of formed sheet panels must preserve geometric manufacturing conditions while exploring layouts and improving performance. STAMP embeds these conditions directly in the shape representation. Each panel is described by its material layout and a reference surface with one height at each occupied position in a plane perpendicular to the drawing direction. Nominal thickness is prescribed along the normal. The representation allows curved surfaces, ribs, and openings. It excludes reference-surface undercuts and keeps nominal thickness fixed without manufacturing penalty terms. Because these conditions belong to the representation, every generated and optimized panel satisfies them regardless of training-set size. Boundary-conditioned diffusion generates alternative layouts, and predictor gradients guide a surface optimization that retains outlines, openings, and designated fixed regions. We demonstrate STAMP on automotive hood inner panels using CarHoods10k. With a comparable number of stored values, depth maps reconstruct surfaces more accurately than point-cloud and voxel baselines. From only five training designs, one of three training runs yields novel and varied layouts. A generator trained on 500 designs, 6.5% of the training set, already reaches the layout variety of the full-data model. Across 11 held-out design families, predictor-based selection among existing dataset designs yields an average gap of 0.18 mm to the best available FEA deformation, versus an expected 1.59 mm for random selection. It does better than random selection in all 11 families. At a 10% depth-change bound, a network not used in the search predicts an average deformation reduction of 24.8% for six generated and existing panels, with a 3.2% average mass increase computed from geometry. Engineers can thus explore and optimize panels within this manufacturing-informed design space.
- [352] arXiv:2611.18145 [pdf, html, other]
-
Title: Compiling Specialized GPU-Initiated NetworkingComments: 20 pages, 13 figures, 6 tables. Equal contribution from first two authors. Code: this https URLSubjects: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG); Programming Languages (cs.PL)
Machine learning systems specialize computation to the workload and hardware, but communication remains in generic libraries whose device code and transport decisions are not visible to high-level compilers. We present Obaleth, a compiler that makes GPU-initiated communication part of a kernel's intermediate representation, co-optimizes it alongside computation, and lowers it to specialized transport code. Obaleth plans completions around buffer lifetimes and optimizes memory ordering, synchronization, and batching across NVIDIA and AMD GPUs. On H100 GPUs, Obaleth sends small messages up to 3.3$\times$ faster, and on 256 GPUs runs all-to-all 1.25$\times$ and 2.3$\times$ faster than NVSHMEM and NCCL GIN, with much smaller memory and register use. In vLLM, it serves Qwen3-235B on 16 B200 GPUs and answers requests up to 1.29$\times$ faster than with DeepEP.
- [353] arXiv:2611.18147 [pdf, html, other]
-
Title: Evidence-Carrying Transport Across EffectsSubjects: Logic in Computer Science (cs.LO)
A program logic proved sound against one effectful semantics need not remain sound when the effect is re-implemented, yet the semantic relationship between reasoning before and after a change of effect is usually left implicit and re-established by hand. This paper provides a uniform, proof-relevant account of it. Working in realizability, where a notion of computation determines a model whose entailments carry explicit program evidence, we introduce simulations between effectful realizability cores, generalizing Longley's applicative morphisms to computation over an arbitrary strong monad. A single simulation transports the source logic along a ladder: it induces a morphism of the associated evidenced frames and, functorially, of the induced triposes. These passages are $2$-fully-faithful, so the three levels agree and a change of effect carries the entire higher-order program logic together with the evidence witnessing each derivation. A computationally dense simulation moreover computes a modality on the source logic, a nucleus through which the transport factors and which marks exactly the reasoning stable under the change, read off from the translation rather than postulated. This recovers known examples such as the double-negation modality of classical realizability, and examples such as sheafification arise from a modal refinement of the construction. Overall, the framework makes a change of effect a first-class object, one that can be composed, factored through its modality, and shown in cases not to exist, opening a path to transporting verified reasoning across effects.
- [354] arXiv:2611.18148 [pdf, html, other]
-
Title: Certificate-Based Verification of Pulsed Field Ablation ModelsComments: 18 pages, 5 figures, 7 tablesSubjects: Computational Engineering, Finance, and Science (cs.CE); Numerical Analysis (math.NA); Medical Physics (physics.med-ph)
Computational models of pulsed field ablation predict the lesion as the tissue where the electric field exceeds a lethal threshold. Their numerical error is rarely bounded, and automated pipelines increasingly run them without an expert checking each result. To verify each run, we separate the question "was the stated model solved correctly?" from the solver that answers it. An untrusted solver returns with its answer a certificate, consisting of a potential with the correct electrode and ground values and an exactly charge-conserving current. A small independent checker verifies the certificate against a written specification of the physics. Using the Prager-Synge hypercircle theorem evaluated in ball arithmetic, it turns the certificate into guaranteed two-sided bounds on the impedance and on the root-mean-square field of every tile of a lesion map, also over a range of blood/tissue conductivity ratios. A vectorized version of the checker takes about a tenth of the solve time. For a focal electrode in contact, the tile-resolved depth at 1000 V and 700 V/cm is proven to lie in 3.50-3.75 mm for every ratio from 1.7 to 3.75. For a bipolar catheter, certified depths overlap published in-vivo depths. Where they differ from a published reference model, the certificates prove that the difference is not numerical error in our computation. We compare the certificate with a grid convergence study on the same solver. On graded grids the study is sharper for local fields but remains an estimate. On uniform grids it fails to converge at the electrode edge, where the certificate still gives a valid bracket. Solvers that deliberately solve equations other than the specified ones pass the grid convergence study, but the certificates refute their claims or leave them unsupported.
- [355] arXiv:2611.18149 [pdf, html, other]
-
Title: Which Model, For Which Variable? Per-Variable Forecast Combination in Sparse ICU Time SeriesComments: Accepted at Machine Learning for Health (ML4H) 2026 Findings Track, SydneySubjects: Machine Learning (cs.LG)
Clinical forecasting papers typically report a single best architecture. We show that on multivariate ICU physiology this question conceals variable-specific differences: no architecture is best on every variable, and which model wins changes from variable to variable. Across five ICU databases (MIMIC-III, MIMIC-IV, eICU, HiRID, and UMM (University Medicine Mannheim), a German single-center ICU) we forecast 94 channels over a 24-hour horizon with a pool of 14 models, and combine them with weights fitted per variable. On test, the combination beats the best single model on all five databases by 6.1-11.0% mean squared error (mean 8.7%). The gain is concentrated where clinical data is thinnest: the sparsest third of channels improves by 11.9% and laboratory values by 7.5%, against 3.1% for the densest third. Rank-based harmonic weighting of the best five models per variable, with no fitted weights at all, recovers about four-fifths of the gain. Extending the pool afterwards with a per-channel gradient-boosted tree, ARIMA and ETS makes the tree the validation-selected best single model on all five databases; per-variable combination still improves on it by 2.9-5.9% (significant on four of five), while equal weighting of the same members does not. As on univariate benchmarks, learned combination beats simple averaging - but only once the weights are constrained. We also report which model to prefer when, and show that the ordering changes with database size.
- [356] arXiv:2611.18152 [pdf, html, other]
-
Title: The Complexity of Traffic Engineering in Segment Routing NetworksSubjects: Data Structures and Algorithms (cs.DS)
Segment routing is a modern variant of source routing which has recently become popular in the Internet. It allows operators to equip network packets with an additional header which specifies segments: a list of waypoints the packet needs to traverse on its way to the destination. This is reminiscent of Valiant routing in parallel computing and introduces an additional degree of freedom in traffic engineering. This flexibility can be exploited to minimize congestion in the network.
This paper sheds light on the computational complexity of traffic engineering in segment routing networks. We consider two problem variants: a fixed-length variant where link weights are given and only waypoints can be chosen, and a variable-length variant where additionally the network operator can also choose the link weights, which then define the shortest paths taken between the waypoints. We prove that the first variant is NP-hard even if the network graph is planar and if all demands and link weights are 1. We also show that the second variant is NP-hard already for planar network graphs with all demands being 1. On the positive side, we present a compact and general mixed integer linear program to compute exact solutions for the first variant. - [357] arXiv:2611.18156 [pdf, html, other]
-
Title: Hamming and Johnson-Type Bounds for Joint Error Correction and DetectionComments: This manuscript is a preliminary draft and is currently being revised. The final version will be posted once the revisions are complete. Comments and suggestions are welcomeSubjects: Information Theory (cs.IT)
We derive upper bounds on the size of codes for joint error correction and detection over classical $q$-ary Hamming spaces. A $(c,c')$ joint error correcting and detecting code corrects every error of weight at most $c$ and detects every error whose weight lies between $c+1$ and $c+c'$. We establish equivalent conditions for this decoding requirement using Hamming balls and Hamming rings, and derive a joint bound that reduces to the Hamming bound when $c'=0$. We introduce joint perfect codes and show that binary repetition codes attain the joint bound whenever $n=2c+c'+1$. For codes satisfying $d_{\min}=2c+c'+1$, we investigate an alternative joint bound based on ring multiplicity and constant-weight code bounds. A further argument using the Cauchy-Schwarz inequality and pairwise ring intersections yields a refined joint bound that is always at least as strong as the original joint bound. We then incorporate an additional outer Hamming shell to obtain a Johnson-type extension. This extension provides a strict improvement for binary codes with $c'=1$, $d_{\min}=2c+2$, $c\geq2$, and $d_{\min}+2\leq n\leq3d_{\min}/2-1$. These results connect sphere-packing bounds with the geometry of joint error correction and detection.
- [358] arXiv:2611.18157 [pdf, other]
-
Title: A Multi-Layered AI Pipeline for BPMN Generation from Natural LanguageSubjects: Computation and Language (cs.CL)
The automated transformation of unstructured natural language process descriptions into Business Process Model and Notation (BPMN) diagrams remains a challenging task, due to linguistic ambiguity and the difficulty to detect non-adjacent control-flow relationships. This paper proposes a solution for generating BPMN models from natural language descriptions through a multi-layered pipeline that integrates transformer-based architectures, Large Language Models and ontology population. The proposed method starts with natural language descriptions and separates textual processing, control-flow relation extraction, and structural validation into dedicated components. Entity extraction is performed using a fine-tuned RoBERTa-based transformer model while a control-flow engine captures non-adjacent relationships through LLM-based inference. To improve generalization under limited available data, the approach uses a dynamic few-shot strategy with con- trolled template input construction, enabling improved generalization under limited data availability. Additionally, an ontology population mechanism is introduced to represent the generated process models as semantic knowledge structures. For evaluation purposes we have used from the literature and industry reference processes. Additionally, we have validated the proposed method on a real-world heart disease patients transitional care process, capturing practical constraints and multi-actor collaboration, and serving as an expert-defined benchmark scenario. Experimental results demonstrate strong performance in the relation extraction task, achieving an accuracy of 90.7%, as well as maintaining structural alignment between the generated and reference BPMN models across multiple evaluation scenarios scenarios.
- [359] arXiv:2611.18163 [pdf, html, other]
-
Title: On the Limits of Sufficient ExplanationsSubjects: Machine Learning (cs.LG)
We formally characterize the conditions under which sufficient explanations exist for classifiers on structured data such as graphs and strings. We focus on post-hoc explainers that select parts of the input as sufficient explanations. An explanation is sufficient if the selected part alone ensures the predicted label. Sufficient explanations have been widely adopted across data modalities, including graphs, strings, and images. However, a unified formal analysis of sufficient explanations remains missing. We bridge this gap with four main contributions: (i) We define a general explanation formalism consisting of an instance space, a classifier, and a select relation. The select relation specifies which potential sub-instances or feature subsets can be selected as explanations from each instance; (ii) We show that the existence of sufficient explainers is fully characterized by the interplay between the classifier and the select relation; (iii) We extend this analysis to practically relevant variants of sufficiency, including probabilistic sufficiency and deletion sufficiency; (iv) Based on the formal analysis, we devise statistics that bound the maximum achievable sufficiency for a given task. We evaluate these statistics on language, protein, and molecular classification datasets and find that full sufficiency is unattainable on most of them. Since these bounds hold for every explainer, they provide a task-level reference against which any explainer's sufficiency can be judged in absolute terms.
- [360] arXiv:2611.18166 [pdf, other]
-
Title: An entropy-stable splitting strategy for a hyperbolic dispersive shallow-water model on unstructured meshesFrédéric Couderc, Arnaud Duran, Maria Kazakova, Bastien Lombard, Pascal Noble, Rémy Baraille, Jean-Paul VilaSubjects: Numerical Analysis (math.NA)
This paper presents an entropy-stable splitting strategy for the numerical solution of a hyperbolic approximation of the Serre--Green--Naghdi (SGN) model, leading to an efficient solver for dispersive water waves on unstructured meshes. The model was introduced by Escalante et al. (2019) and justified shortly thereafter by Richard (2021) as an asymptotic model of the free-surface Euler equations under a weak compressibility assumption. While the parent model includes full topography terms and enhanced dispersive properties, the present study focuses on its mild-slope form without enhanced dispersion, which retains an exact energy equation, the cornerstone of the stability analysis. The splitting yields a shallow-water subsystem with two passive scalars, governed by the \textit{slow} gravity waves, and an acoustic subsystem carrying the dispersive effects, governed by the \textit{fast} acoustic waves. The gravity step adapts an existing entropy-stable scheme, while an original scheme is derived for the acoustic step; both dissipate the energy at the discrete level, under local fully discrete stability criteria on general unstructured meshes. Complementary linear analyses select the low-dissipation calibration of the gravity step used in the computations, outside the scope of the nonlinear estimate. The acoustic step, which requires no global linear solve, is sub-iterated after each gravity time step; the resulting solver is three to six times more expensive than a classical shallow-water solver, depending on the chosen acoustic speed. Numerical validations, ranging from academic benchmarks to a realistic coastal configuration, highlight the properties of the approach and pave the way for operational use.
- [361] arXiv:2611.18168 [pdf, html, other]
-
Title: Who Counts as Exposed? Measurement Choices in an Open, Re-estimable Index of Occupational AI ExposureSubjects: Information Retrieval (cs.IR)
Occupational AI exposure indices are used as inputs to economic analysis, workforce planning, and policy, and the choices built into them decide which workers are described as exposed. We build an open, re-estimable task-level index and vary its two main design decisions, context and calibration, to show how much the answer depends on them. For each of the 18,796 occupation--task pairs in \onet, seven open-weight reasoning models rate exposure under a common rubric. Each model rates with or without a context of retrieved news and research passages, workplace descriptors, and a bottleneck-based reasoning protocol, and the ensemble is combined with or without simplex-constrained calibration to observed Claude usage. The largest difference from the dominant 2023 benchmark is not a disagreement about what AI can do but about how much credit to give exposure that depends on software integrations not yet built, a choice that no data can settle. At the benchmark's own weight, the calibrated index counts fewer exposed workers than the benchmark, and when only direct exposure counts it counts more. At full score resolution, the ensemble ranks tasks better than the benchmark against observed usage and an independent physical-task classification. Adding context mainly shifts the level of exposure rather than improving the ranking. Context and calibration both relocate exposure toward higher-wage occupations, and calibration, which anchors the index to one platform's users, does more of it, so we also release the uncalibrated arm. Exposure indices used as policy inputs should state such choices, report their numbers under the alternatives, and be re-estimated as systems change.
- [362] arXiv:2611.18169 [pdf, html, other]
-
Title: The QBF Gallery 2026Comments: Open access versionSubjects: Logic in Computer Science (cs.LO)
The QBF Gallery 2026 is the latest evaluation event of (dependency) quantified Boolean formula (QBF) solving technology. Its goal is to assess the current state-of-the-art, enable the evaluation of novel solvers on newly curated benchmark instances and present the progress made since the previous competition. This report discusses the solvers, instances and tools submitted by the community and introduces new benchmark collections for evaluation. We evaluate solvers in three tracks, reflecting most common input formats in (dependency) QBF solving. In addition to those main evaluations, we investigate effects of preprocessing and include a track of new crafted benchmark families.
The results are analyzed using various metrics and complementary statistics on solver similarities, different memory limitations and performance across formula families. Finally, the dedicated benchmarks for each track are made available as an artifact to support future research and reproducible evaluation. - [363] arXiv:2611.18170 [pdf, html, other]
-
Title: ALoMa: Affine Correspondences Beyond PatchesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Affine correspondences (ACs) enable robust and efficient solvers for multiple-view geometry by constraining the local geometry of point correspondences (PCs). Despite their utility in minimal solvers, AC estimation remains understudied. We revisit AC estimation beyond local image patches, extending it to the image and image-pair levels. Rather than estimating the components of the affine transformation separately, we learn the full 4-DoF matrix jointly. Building on this formulation, we introduce ALoMa, an AC estimator using LoMa keypoint descriptions, demonstrating consistent improvements across a wide range of downstream tasks. Code at this https URL.
- [364] arXiv:2611.18172 [pdf, html, other]
-
Title: STaG: Human-Annotation-Free Unsupervised Domain Adaptation for Text-Based Person RetrievalSubjects: Computer Vision and Pattern Recognition (cs.CV)
Real-world deployment of Text-Based Person Retrieval (TBPR) is bottlenecked by the absence of textual annotations in new domains. Existing Unsupervised Domain Adaptation (UDA-TBPR) methods still assume access to target-domain human-written annotations (albeit unpaired with images), limiting their practicality. We consider a more practical setting: human-annotation-free UDA-TBPR, with labeled image-text data in the source domain and annotation-free, unlabeled images in the target domain. This setting introduces two core challenges: the absence of cross-modal alignment in the target domain, and visual domain shifts from source to target. To tackle these, we propose STaG, a dual-path framework comprising Style-Agnostic Text Alignment (SATA) and Heterogeneous Graph Memory Alignment (HGMA). Specifically, SATA leverages source-domain data to decouple identity-discriminative semantics from machine-generated stylistic artifacts, establishing robust cross-modal alignment for the target domain. Meanwhile, HGMA exploits cross-domain semantic neighborhoods via a memory-based graph to learn domain-invariant visual features. Ultimately, STaG achieves state-of-the-art performance across three benchmarks under this realistic setting.
- [365] arXiv:2611.18175 [pdf, html, other]
-
Title: Safety Analysis of Parameterized Networks with Shared CountersSubjects: Logic in Computer Science (cs.LO)
We study networks with an unbounded number of processes sharing some natural counters that can be decremented, incremented and tested to zero. We look at two safety problems: the first one asks whether there is a network with an execution bringing a process in a given final state and the other one asks the execution to bring all the processes in the final state. In their full generality, these two problems are undecidable, but we show that decidability can be regained by either using a single counter or by forbidding the use of zero-tests (as it is is the case in Vector Addition Systems with States). For these decidable cases, we obtain low complexity results, the problems being solvable either in non-deterministic logarithmic space or in polynomial time.
- [366] arXiv:2611.18176 [pdf, html, other]
-
Title: ESI: An Embodied Self-Improvement Flywheel for LLM-VLA AgentsComments: Technical report. 25 pages, 7 figures. Code: this https URLSubjects: Robotics (cs.RO)
Independently pretrained language models and vision-language-action (VLA) policies provide complementary building blocks for embodied agents, but connecting them does not automatically produce coordinated behavior. The VLA may generate actions that diverge from the LLM's intended subtask and, upon regaining control after a correction, undo the promising state restored by the LLM. We introduce ESI (Embodied Self-Improvement), an automated flywheel that uses environmental interaction to align reasoning and action. ESI treats the modular agent as the unit of post-training, repeatedly refining the prompts that guide the LLM, the shared skill interface that expresses its intent, and the VLA parameters that realize that intent, then redeploying these components together. Our key insight is that collaboration itself can supply supervision for improving this interface. The LLM's subtask decisions provide semantic supervision for what the policy should execute, while its effective corrections and subsequent student continuations provide action-level supervision for recovery and handoff. The student also learns from its own effective execution, allowing locally useful experience to contribute even when the overall episode fails. On ten selected RoboDojo tasks, ESI achieves a mean Score of 62.03 and a mean success rate of 49.41% using an LLM-VLA agent without generating task-specific control code. On both metrics, it outperforms the published reference results of all hybrid embodied agents, including the Agent+WAM system GO-3 Nano Preview (47.27 Score and 36.60% success), as well as all compared code-based methods.
- [367] arXiv:2611.18179 [pdf, html, other]
-
Title: Explanations that Transfer: Interpreting Medical Retrieval with Sparse AutoEncodersComments: Proceedings of Machine Learning Research 346, 2026 (Machine Learning for Health, ML4H 2026). Code: this https URLSubjects: Information Retrieval (cs.IR); Machine Learning (cs.LG)
Medical AI systems that answer users' queries rely on retrieving medical knowledge from large document collections. In this context, finetuning general models on medical datasets can improve retrieval relevance. Yet the dense embeddings underlying modern retrieval systems remain largely opaque, making it difficult to understand the mechanics behind their results. Moreover, adapting to a specific medical dataset risks losing transferability to new clinical corpora. In this work, we tackle these issues by applying Sparse Autoencoders (SAEs) to medical information retrieval, transforming dense text embeddings into semantically interpretable representations. To do so, our model combines a reconstruction objective with an InfoNCE contrastive loss that directly preserves query-document relevance. We evaluate three embedding models across six medical retrieval datasets, both in-domain and out-of-domain. Trained on a medical corpus, our sparse codes improve in-domain retrieval over the original embeddings, and lose less out-of-domain performance than their dense counterparts, especially when the target corpus is distant from the training one. Importantly, our experiments show that the method yields interpretable features whose semantic content can be inferred from the queries and documents that activate them most strongly. The interpretation of these features sheds light on why a particular document was retrieved, and transfers beyond the training distribution.
- [368] arXiv:2611.18183 [pdf, html, other]
-
Title: When Should a Virtual Patient Ask? RISP-Guided Proactive Information Seeking in Multi-Agent Patient SimulationJingyao Cai, Shuaijun Liu, Abdul Rehman, Xiaowan Li, Qintian, Thomas Dolby, Sue M Green, Xiaosong YangSubjects: Multiagent Systems (cs.MA)
Virtual patient (VP) systems support nursing communication training by providing simulated encounters in which learners practise explaining care and responding to patient concerns. Although large language models (LLMs) support flexible dialogue, current approaches give less explicit attention to how patients' information needs develop and guide the content and timing of their questions. We draw on the Risk Information Seeking and Processing (RISP) model, which relates information seeking to perceived risk and the gap between perceived knowledge and desired understanding. Guided by this perspective, we propose the RISP Multi-Agent System (RISP-MAS), a VP system for patient-initiated information seeking. The system links ongoing assessment of the patient's perceptions and information needs to coordinated decisions about whether to seek information, when to ask and which uncertainty to address. Evaluation of nursing student-VP interactions suggested that RISP-MAS exhibited plausible patient information-seeking behaviour, as indicated by expert assessments, while student feedback reflected positive interaction experiences. This work extends patient simulation by connecting a psychological account of information needs to explicit decisions about question initiation. The approach offers nursing students opportunities to practise recognising unresolved patient concerns and tailoring explanations to patients' information needs.
- [369] arXiv:2611.18184 [pdf, html, other]
-
Title: PTMTD: A Topology-Aware Temporal-Graph Attention Network for Joint Cyber-Attack Detection and Relay-Level Localization in Power System Protection NetworksSubjects: Systems and Control (eess.SY); Signal Processing (eess.SP)
The digitization of protection relays exposes power grids to sophisticated cyberattacks, yet existing GNN detectors omit temporal encoding, rely on static topologies, and lack relay localization. We propose PTMTD-a Pearson-driven Temporal Multimodal Topology-aware Detection Network-coupling a dilated TCN with multi-head GATv2 in a dual-head architecture for simultaneous 37-class detection and four-relay localization. PTMTD constructs its graph from measurement correlations (no impedance data) and is trained with a composite loss jointly optimizing both tasks. On the MSU-ORNL benchmark (78,377 samples, 37 classes), it achieves 99.27% accuracy, 99.25% macro F1, and 95.98% localization, outperforming GraphKAN, LSTM, GCN, and Random Forest by 5.64-11.15 percentage points (p < 0.05). Ablations confirm gains from temporal encoding (+2.38%), graph attention (+2.69%), residuals (+2.07%), and correlation-driven supervision. With 639k parameters and 2.5 ms CPU inference, PTMTD is lightweight and deployment-ready. While validated on a transmission testbed, the architecture applies to microgrids with similar relay structures. Code and models are publicly available.
- [370] arXiv:2611.18186 [pdf, html, other]
-
Title: NPU-Accelerated GNSS Signal AcquisitionSubjects: Robotics (cs.RO)
On small single-board computers (SBCs), GNSS signal acquisition is a major CPU load. This paper proposes using the neural processing unit (NPU) of an SBC not for machine learning but as an accelerator for the parallel code search in GNSS signal acquisition. The discrete Fourier transform (DFT) used in acquisition is decomposed by the four-step DFT into matrix products of small DFTs and expressed by $1\times1$ convolutions, element-wise products, and transposes, which NPUs execute efficiently. Furthermore, we propose a method that divides the DFT, the non-coherent integration, and the peak detection between the CPU and the NPU according to the configuration of the NPU (an NPU with its own memory connected via PCIe, or an NPU sharing memory with the CPU). The proposed method is implemented in the open-source software GNSS receiver Pocket SDR and evaluated on a Raspberry Pi 5 combined with the external AX8850 NPU and on the NPU built into the SoC (RK3588) of the ROCK 5C, a small SBC of comparable size. Compared with acquisition by an optimized CPU implementation, the acquisition of short codes on the Raspberry Pi 5 with the AX8850 is 3.0-3.8 times faster. In the evaluation of the whole receiver, the number of tracking channels that reach their first lock within 30 s with the NPU is 1.5 times that with the CPU, and the time until 90% of the channels are locked is reduced from 66 s to 26 s. Furthermore, because the CPU load of acquisition is reduced, the number of tracking channels that can be processed in real time increases from about 175 to about 230. These results show that using the NPU on SBCs with limited CPU resources is effective for achieving both GNSS signal acquisition and tracking. The implementation is publicly available as open source under the name PocketSDR-NPU.
- [371] arXiv:2611.18189 [pdf, html, other]
-
Title: SemRAG: Semantic-Guided Sentence-Level Retrieval-Augmented Sign Language ProductionSubjects: Computer Vision and Pattern Recognition (cs.CV)
Sign language production (SLP) aims to generate sign pose sequences that faithfully express spoken-language text while preserving the rhythm and coordination patterns of signing. Common SLP methods learn text-to-pose mappings from training data but often suffer from semantic inaccuracies in long or complex sentences. Retrieval augmentation offers a way to introduce realistic signing priors to provide more explicit guidance for text-to-pose generation. However, existing retrieval-based SLP methods typically operate at the gloss or word level, relying on gloss-pose dictionaries or isolated sign segments that require costly annotations and fail to preserve continuous sentence-level dynamics. To bridge this gap, we propose SemRAG, a semantic-guided sentence-level retrieval-augmented framework that retrieves full-sentence signing templates as priors without constructing gloss/word-level sign dictionaries. While such templates preserve local motion patterns and global rhythm, they also introduce a semantic mismatch problem: irrelevant motions in the retrieved sentence may interfere with the target semantics. To address this issue, we introduce an SLT-guided template semantic localization module that identifies target-relevant temporal regions within the retrieved template while suppressing irrelevant signing content. To adaptively incorporate the template features into the generator, we further introduce layer-specific learnable gates that regulate their contribution across generator layers. Extensive experiments show that SemRAG improves semantic consistency and motion quality over baselines under both back-translation and motion accuracy metrics.
- [372] arXiv:2611.18190 [pdf, html, other]
-
Title: MarathonTS: a benchmark of long synthetic time series from window assembly and long-series generatorsSubjects: Machine Learning (cs.LG)
Many applications of synthetic time series require a series that is much longer than the windows on which a generative model is trained. Such a series can be assembled from a pool of short generated windows, which requires deciding which windows are taken, in what order, and how their seams are processed, or it can be produced directly by a long-series generator. This study presents MarathonTS, to our knowledge the first open benchmark that compares both families under one protocol. Every method produces a series of the same length on 9 open hourly datasets with 10 seeds. The series is evaluated by its fidelity to the real data (83 metrics in 12 groups, against three real references: train, validation and an out-of-time split), by its utility (models are trained on the synthetic series only and tested on real data, TSTR against TRTR, in 14 classification and regression tasks), by its memorisation of the training data and by its computational cost. The statistical analysis is carried out as a separate stage that reads the published tables and gives the rankings of every method in every slice, the answers to the research questions with tests and uncertainty, and the results of a pre-specified confirmatory family. In the main run, the methods differ in both confirmatory cells; the choice of windows matters more than the processing of their seams, which does not improve the series, and no method trains the downstream models as well as the real series. The top tier of the leaderboard is reproduced on other hardware, whereas the exact places within it are not. For the reproducibility of the study, the code, the data hashes and all results are publicly available (links below).
- [373] arXiv:2611.18196 [pdf, html, other]
-
Title: Movable Antenna-Aided ISAC: A Time- and Frequency-Division FrameworkComments: 13 pages, 9 figures, 1 tableSubjects: Information Theory (cs.IT)
This paper investigates movable-antenna (MA)-enhanced integrated sensing and communication (ISAC) under time division duplex (TDD) and frequency division duplex (FDD) protocols, where sensing and communications occupy orthogonal time and frequency resources, respectively. A wideband target-position Cramér-Rao bound (CRB) is derived under a pulse-radar framework with pulse compression, incorporating delay and angle estimation error bounds in Cartesian coordinates. Both bounds depend on sensing power; the delay bound also depends on sensing bandwidth, while the angle bound depends on the effective aperture determined by the antenna position vector (APV). A worst-case position CRB over a prescribed target region ensures robust sensing performance. For multiuser downlink communication, joint APV and resource-allocation problems are formulated for TDD and FDD to maximize the minimum achievable user rate subject to the sensing requirement. An alternating optimization algorithm is developed for FDD, while TDD is addressed by searching over feasible sensing pulse numbers combined with APV optimization. Representative scenarios clarify how MA improves channel orthogonality and enlarges the sensing aperture, and how these effects shape the TDD/FDD rate comparison. Numerical results show that APV optimization enlarges the communication-sensing tradeoff region over fixed-position antenna benchmarks. The optimized configurations exhibit distinct tendencies: TDD favors a larger antenna position variance, whereas FDD favors better channel orthogonality among users.
- [374] arXiv:2611.18197 [pdf, html, other]
-
Title: Data-Driven Sparse Identification and Robust Tube-Based Predictive Control for Multivariable Chaotic SystemsSubjects: Systems and Control (eess.SY)
This paper addresses the data-driven synchronization and robust control of multivariable chaotic systems subject to strict actuation constraints, modeling uncertainties and bounded input disturbances. Using empirical data collected from a four-dimensional chaotic Rossler response system, Sequentially Thresholded Least Squares (STLSQ) sparse regression is employed to recover a nominal control-affine model, achieving an identification accuracy of R^2 = 1.0000 across all channels with root-mean-square error within [4.15e-6, 8.33e-4]. To establish a mathematically rigorous bridge between data-driven identification and robust synthesis, all structural discrepancies in the drift vector field and input matrix are unified into a single lumped uncertainty field delta(x,u). A computable scalar bound delta_bar is established based on library truncation error, regression residuals, input bounds, and Lipschitz constants (Lemma 1). We subsequently prove that input-output feedback linearization maps delta(x,u) into a state- and control-dependent additive disturbance d(z,nu) in Brunovsky canonical coordinates, bounded by ||d(z,nu)||_2 <= alpha_1 delta_bar + alpha_2 delta_bar ||nu||_2 with closed-form expressions for alpha_1 and alpha_2 (Theorem 1). This certified bound, together with a bounded input disturbance and a zero-order-hold margin, yields a compact disturbance set that seeds a robust tube-based model predictive controller. We prove that the resulting tightened quadratic program is recursively feasible, that the physical actuator limit (|u_i| <= 1100) remains inactive under an explicit sufficient condition, and that the closed-loop tracking error is Input-to-State Stable (ISS) with an asymptotic gain that is linear in delta_bar plus a term due to the exogenous input disturbance (Theorems 2 and 3).
- [375] arXiv:2611.18199 [pdf, html, other]
-
Title: T2ST: Data-Efficient Spatio-Temporal Foundation Models via Temporal Data and Knowledge TransferSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Spatio-temporal foundation models (STFMs) require large-scale training corpora, yet synchronized observations and spatial metadata are costly to acquire and curate, limiting the availability of suitable data. By contrast, time-series data are abundant and easier to collect. Since observations at each location form a time series, this motivates a natural question: can abundant time-series data enhance STFMs when spatio-temporal observations are scarce? To bridge this gap, we propose T2ST, a data-efficient training framework that augments limited spatio-temporal observations with selected auxiliary time series and transfers knowledge from a frozen time-series foundation model. Our key insight is that temporal pattern learning can benefit from abundant external time series, while spatial dependency learning remains grounded in real spatio-temporal observations. Guided by this insight, T2ST retrieves matched sequences to reinforce sparsely observed temporal patterns and selects reliable complementary sequences to broaden temporal coverage. Temporal relation distillation aligns relationships among historical patches, while multi-scale prediction distillation transfers forecasting behavior at fine and coarse temporal resolutions, enabling knowledge transfer across heterogeneous architectures. Using only 10\% of the original spatio-temporal data, T2ST enables STFMs to achieve overall forecasting performance competitive with or superior to their full-data counterparts across six long-term and twelve short-term forecasting datasets. These results establish a new paradigm for data-efficient training of spatio-temporal foundation models. Code is available at this https URL.
- [376] arXiv:2611.18201 [pdf, html, other]
-
Title: Go under Chinese Rules with a Succinct Game History is Complete for Exponential SpaceComments: A preliminary version of this paper appears in the Proceedings of the ACM-SIAM Symposium on Discrete Algorithms (SODA 2027)Subjects: Computer Science and Game Theory (cs.GT); Computational Complexity (cs.CC)
The complexity of games is a central topic in combinatorics, algorithms, and computational complexity theory. It asks how difficult it is to determine the winner of a combinatorial game under optimal play. Many games, such as Hex or Othello, are PSPACE-complete: the game tree has polynomial depth, so the game can be decided in PSPACE. However, some games allow exponentially long plays, for instance Chess (without a generalized 50-move rule) or Go under Japanese rules; these are complete for exponential time. To prevent infinite games, some games employ a no-repetition rule (super-ko rule), for instance Go under Chinese rules. Robson (1984) proved that Chess and Checkers with a no-repetition rule are even harder: deciding whether the first player has a winning strategy is EXPSPACE-complete. Since then, it has been an open problem whether Go with a no-repetition rule is EXPSPACE-complete, too.
The complexity of a game under a no-repetition rule is typically interpreted as follows: given a board position, decide whether the first player has a winning strategy such that no repetition occurs from that position on. We introduce the concept of a succinct game history: a Boolean circuit that, given a board position, outputs 1 if the position is in the game history and 0 otherwise. It serves as a "judge" that tells the players whether a move is illegal because it creates a repetition. We prove that deciding whether Black has a winning strategy in Go with a no-repetition rule is EXPSPACE-complete when the game history is given succinctly. Moreover, the starting position in our hardness proof is reachable by legal gameplay, which has not always been the case in prior proofs. - [377] arXiv:2611.18213 [pdf, html, other]
-
Title: Who sets the consent signal? Platform defaults and crawl coverage in measuring opt-outs from AI trainingComments: 24 pages, 6 figures, 2 tables. Dataset: this https URLSubjects: Networking and Internet Architecture (cs.NI)
Web data used in social research and in the training of AI models increasingly carry consent signals, but who sets these signals is rarely measured. This paper attributes opt-outs from AI training to the party that set them. The complete lists of Polish, Czech, Slovak, Hungarian and German domains recorded by Common Crawl, with 3,445,751 reachable sites, were crawled in October 2026 together with a random sample of 10,000 Slovak registry domains. Each signal was matched against vendor templates and identical files. Four exploratory hypotheses were examined with logistic models. Between 0.70% and 3.82% of the listed sites opt out. Platform and hosting defaults set 15% to 78% of the opt-outs. Settings active by default reach 60% to 95% of a vendor's sites, while settings that owners switch on reach at most 2.1%. Most of the higher Czech and Slovak rates compared with Germany are statistically attributable to platforms. Lists derived from Common Crawl miss about a third of the Slovak opt-outs because they exclude sites that block its crawler. The results describe delegated consent, a signal that a platform sets on behalf of its users, which measurements of consent on the web must consider.
- [378] arXiv:2611.18219 [pdf, html, other]
-
Title: Large language models put existential risk from artificial intelligence at 10%Subjects: Computers and Society (cs.CY)
How likely is it that advanced artificial intelligence (AI) will cause human extinction or a similarly permanent disempowerment of humanity? Humans, including AI experts, frequently disagree on this question. Yet one important perspective is missing from this debate: that of AI systems themselves. We put the question to 241 large language models (LLMs) at every available reasoning setting. The models converged on a median risk of 10.0%, with the middle 90% of their estimates between 3.5% and 15.8%, far narrower than the spread among human experts. More expensive, and typically more capable, models stated higher risks; newer models showed more uniform, but not higher, risks. More reasoning barely changed the models' estimates but widened their uncertainty, mostly toward higher risk. Almost all models named loss of control over AI as the most likely cause. These answers are not independent judgments; they likely echo training data, including published expert polls, and choices made in aligning the models. Still, they document the position that current AI systems state on the risks of their own technology. Whether this position reflects any genuine understanding of the systems' own capabilities remains open. Repeated measurements can show whether it shifts as AI systems advance, and how training and alignment shape it is a question for both research and public debate.
- [379] arXiv:2611.18220 [pdf, html, other]
-
Title: From Pen-and-Paper to Machine-Checked Proofs: Lean for System IdentificationSubjects: Systems and Control (eess.SY); Artificial Intelligence (cs.AI)
Mathematical proofs in scientific papers are often long and intricate, relying on chains of references to earlier results. This makes it difficult for readers and reviewers to verify each and every step. Interactive theorem proving offers a complement: the theorem and its proof are written as code, and every inference is checked by a small, trusted piece of software called the kernel. This paper serves as an introduction to the Lean 4 proof assistant for a system identification and control audience, assuming no prior exposure to formal methods. We first show how a Lean proof is read using two elementary examples: the first on identifiability, the second on set invariance. We then illustrate how Lean can be used to analyse results from the literature, taking as example a classic lemma on the inversion of the noise filter in prediction-error identification (Ljung, 1999, Lemma 3.1). Since in Lean every step must be justified, the formalization brings to light a subtle point: the original claim of that lemma holds almost surely, although not for every realisation, as its statement may suggest. A simple counterexample makes this concrete. Finally, the paper discusses guidelines and perspectives on the use of Lean in system identification research.
- [380] arXiv:2611.18223 [pdf, html, other]
-
Title: Characterisation of the DQsym Framework Against Harmonic State-Space and Dynamic Phasor Models for Modular Multilevel ConvertersComments: Submitted to the 25th IET International Conference on AC and DC Power TransmissionSubjects: Systems and Control (eess.SY)
Modern power systems rely on power converters that exhibit multifrequency periodic dynamics, such as modular multilevel converters (MMCs). Such converters are commonly analysed using harmonic state-space (HSS) and dynamic phasor (DP) representations. Both approaches rely on building state matrices based on the system topology, whose size grows with the number of harmonics retained in the model. This paper shows that a symmetrical direct-quadrature (dq) dynamic phasor formulation for time-domain simulation, named DQsym, has a smaller solver state matrix than HSS and DP, and whose size is independent of the truncation order K. This reduction results from evaluating harmonic coupling outside the state matrix, in the nonlinear signal blocks. The open-loop MMC model used in this paper is derived in three harmonic-domain (HD) formulations and validated against a three-phase time-domain reference model. The models are compared in terms of state-space size and sparsity, harmonic content, steady-state accuracy and eigenvalues. For truncation at K = 4, the DQsym solver has 45 states against 135 for HSS and DP, stays within about 4% of the reference, and, once closed with its coupling blocks, yields eigenvalues identical to those of HSS and DP.
- [381] arXiv:2611.18228 [pdf, html, other]
-
Title: New Results on the Subfield MetricSubjects: Information Theory (cs.IT)
In this paper, we further develop the theory of the recently introduced $\lambda$-subfield metric. We investigate extremal codes, classical bounds and asymptotic densities in this setting.
We first study codes attaining a Singleton-type bound, called Maximum $\lambda$-subfield Distance (M$\lambda$D) codes. We establish structural restrictions on their parameters and extend the theory to sublinear codes. We then investigate the asymptotic density of M$\lambda$D codes in several parameter regimes and compare their behavior with that of MDS and MRD codes.
We then turn to a Plotkin-like bound, where we construct a family of constant-weight codes attaining the bound and classify all non-degenerate linear constant-weight codes for $\lambda>1$.
Finally, we complement the collection of classical coding-theoretic bounds available for the $\lambda$-subfield metric by deriving Griesmer-, Elias- and Johnson-type bounds. - [382] arXiv:2611.18229 [pdf, html, other]
-
Title: A Reinforcement Learning Approach to Space-Filling Input Design for Nonlinear System IdentificationSubjects: Systems and Control (eess.SY)
In this paper, we propose a reinforcement learning (RL) approach to space-filling input design for nonlinear system identification, in which an agent learns a control policy that actively drives the system to explore its state space. Classical input design, developed mainly for linear models, minimises the covariance of the estimated parameters. For nonlinear black-box models, however, the dominant error is the bias induced by the model structure over the region of interest, which motivates space-filling strategies that seek to cover the system state space. Existing space-filling approaches rely on predefined input signal classes and require solving explicit optimisation problems. Unlike these methods, the proposed approach requires no input class to be selected and no frequency band to be tuned for the system at hand, satisfies an amplitude bound by construction, and adapts to varying initial conditions through its observation-based policy without re-optimisation. The learned excitation signals exhibit physically meaningful behaviours, such as resonance excitation and amplitude modulation. The approach is evaluated against tuned swept sine and crest-factor-optimised multisine excitations, and against a state-of-the-art space-filling multisine design, on two nonlinear benchmark systems: a mass-spring-damper with nonlinear stiffness and a Van der Pol oscillator. Under the same amplitude bound, the learned policy achieves the best overall space-filling performance on both systems and distributes its samples more uniformly than the state-of-the-art design.
- [383] arXiv:2611.18230 [pdf, other]
-
Title: Personalization of Basic and Moral Values Guiding LLM-Controlled Assistant RobotsComments: 18 pages, 6 figures. Version accepted for publication in: Value Engineering in Artificial Intelligence: Fourth Inter-national Workshop, VALE 2026, Bremen, Germany, August 15, 2026, Proceedings. Work-shop affiliated with 35th International Joint Conference on Artificial Intelligence (IJCAI-ECAI 2026)Subjects: Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI)
Personal assistant robots powered by large language models (LLMs) are likely to enter everyday life within the next decade, but the values that should guide their behaviour remain largely unspecified. Current alignment practice steers a single model toward an aggregate of human values, ignoring that individual users may prefer a robot whose value profile differs from their own. We administer parallel self- and robot-targeted versions of three established psycho-metric instruments - the simplified Schwartz Values Survey (SVS), the Portrait Values Questionnaire (PVQ-21), and the Moral Foundations Questionnaire (MFQ) - to potential future users (N=22), capturing both their personal values and the values they would want an assistant robot to embody. We compare the three instruments as personalization tools using the UEQ-S and benchmark fifteen contemporary LLMs against the resulting profiles. Participants wanted their robot to score markedly lower than themselves on personal-focus values and higher on social-focus values; moral profiles were nearly identical. LLMs prompted to act as robots shifted only weakly toward the desired profile, indicating that default model behaviour and prompt engineering are no substitute for explicit value personalization.
- [384] arXiv:2611.18235 [pdf, html, other]
-
Title: NeoResearch: Literature- and Experience-Guided Autonomous Time Series Model DiscoverySubjects: Machine Learning (cs.LG)
Developing effective forecasting models for diverse time-series domains requires substantial expert knowledge, evidence from scientific literature on the web, and iterative experimentation. However, translating this knowledge and research experience into model designs requires preserving source context and testing applicability under task-specific assumptions. In this paper, we propose NeoResearch, a literature- and experience-guided autonomous research framework that leverages large language models (LLMs) as researchers to discover lightweight, domain-specific forecasting models. Specifically, NeoResearch establishes an expert-inspired dual-loop research workflow, coupling hypothesis-driven architecture exploration with iterative training and hyperparameter optimization through experimental feedback. To enable knowledge-grounded discovery, the agent autonomously mines scientific literature from the Web, extracting modeling mechanisms, empirical evidence, and applicability conditions to formulate task-grounded, testable architecture hypotheses. Furthermore, a hierarchical experience memory links literature findings to experimental outcomes and consolidates conditional lessons, guiding architecture discovery and training optimization while supporting reuse across forecasting tasks. Experiments on ten datasets show competitive performance and support dual loops, literature extraction, and memory reuse. Our code is available.
- [385] arXiv:2611.18238 [pdf, html, other]
-
Title: Small Is Not Dead: Distance-Aware Cluster Revival for e-meansSubjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
e-means (Bessa et al., 2026) is a recent mini-batch k-means algorithm, now part of the FAISS library, that reaches the error of Lloyd's algorithm about an order of magnitude faster on large embedding datasets. Every five mini-batches, it removes every cluster whose running point count falls below a tenth of the average and reuses its centre to split a large cluster. In the authors' ablation, removing this mechanism, called dead-cluster revival, raises the error more than removing any other component added to the moving-average update. The rule, however, considers only cluster size: it cannot distinguish a cluster that is small because it is redundant from one that covers a rare, isolated group of points, and on data with such groups it repeatedly removes centres whose removal is costly. We add a second condition, based on the distance between centres: a small cluster is removed only if its centre is close to another centre, relative to a running estimate of the mean squared error. When a centre equals the mean of its points, this squared distance is an upper bound on the average increase in squared error, per point of the cluster, caused by removing the centre. The condition costs one nearest-centre search for the small clusters every five mini-batches and adds no measurable running time. On the e-means paper's embedding benchmarks, it never prevents a removal, so the results are identical to those of FAISS. On datasets with small isolated groups not used during development, it reduces the test error after two epochs by roughly 10% to 80% relative to count-based revival, and it preserves the advantage of k-means++ initialization; on the other evaluation datasets, it stays within about 1% of the better of revival enabled and disabled. We also report two failure modes that the analysis predicts. This is a CPU proof of concept; GPU implementations are left to future work.
- [386] arXiv:2611.18239 [pdf, html, other]
-
Title: Neural Classifiers as Error-Correcting CodesComments: 40 pages, 6 figures. Submitted version. Published in Information Sciences 766 (2027) 124394Journal-ref: Information Sciences 766 (2027) 124394Subjects: Machine Learning (cs.LG); Information Theory (cs.IT)
A unified framework is proposed that characterizes supervised deep classification as the decoding of error-correcting codes in the Shannon sense. Cross-entropy optimization is shown to align softmax classifiers with maximum-likelihood decoding, where logits map to Hamming distances to class-specific codewords. This correspondence explains the emergence of standard divergence objectives (e.g., Kullback-Leibler, Jeffreys) and provides a geometric basis for the stability of decision boundaries. Furthermore, the geometries of Neural Collapse and LeJEPA are identified as optimal constellations for decoding under isotropic noise. Unlike classical Error-Correcting Output Codes, deep networks are shown to learn effective codebooks in the feature space implicitly. This insight is leveraged to derive novel criteria for lossless quantization, improved knowledge distillation, and architectural regularization based on code redundancy. Finally, this perspective is extended to self-supervised encoders, offering a cohesive explanation for the geometry of latent representations in modern neural models.
- [387] arXiv:2611.18240 [pdf, html, other]
-
Title: SCAR: Bridging Coarse Retrieval and Fine Localization through Candidate RerankingSubjects: Computer Vision and Pattern Recognition (cs.CV)
Natural-language point-cloud localization typically retrieves candidate map cells and then predicts a position inside each candidate, yet the final output is still determined by retrieval order. A retrieval score measures region-description relevance, not whether the coordinates a region yields are accurate: a relevant region may produce a displaced prediction, and a lower-ranked region may contain a successful one. We study the final selection problem after the candidate pool and Fine predictions are fixed---how to choose the coordinate most likely to localize successfully, and under which conditions the gain holds. We propose SCAR (Spatial Consistency-Aware Reranking). SCAR treats each cell--Fine-coordinate pair as a localization hypothesis, interprets directional relations as functions of position, and reads spatial support at the backbone's actual Fine point (Fine-conditioned Evidence Readout). It organizes this readout with center support, semantic presence, retrieval score, and geometric offsets into a 23-D representation, trains a shared scorer with localization-success labels, and at inference reranks existing hypotheses and returns their original coordinates. Experiments cover four public backbones and a corrected Des4Pos reimplementation across five domains. Excluding one protocol-reference configuration, exploratory TEST point estimates on the remaining 24 configurations show Top-1@5\,m above Original on 23 and Cell R@1 above Original on all 24. On development-set VAL, the Fine readout improves Top-1@5\,m by 5.77 percentage points (pp) over a center surrogate on corrected Des4Pos*KITTI, and the complete scoring scheme improves it by 13.98\,pp over FSR on MambaPlace$\times$nuScenesFull; the latter measures the complete scheme, not learned ranking alone.
- [388] arXiv:2611.18244 [pdf, html, other]
-
Title: RACE: Ranking-Uncertainty Aware Capacity-Adaptive Customization for Efficient Sequential RecommendationTianqi Liu, Kaipu Sun, Kairui Fu, Wenkai Wang, Sihao Liu, Hui Zang, Ran Chen, Tao Wang, Shengyu ZhangComments: Accepted by WSDM 2027 (Main Conference)Subjects: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
Efficient sequential recommendation has emerged as a promising paradigm for providing real-time personalized service. To deploy models under resource-constrained environment, model compression becomes a necessity, yet existing methods typically apply uniform compression ratios, overlooking the heterogeneity in user intent complexity. This "one-size-fits-all" approach leads to a fundamental Capacity-Complexity Mismatch, where computational resources are either wasted on simple instances or insufficient for complex ones. To address this, we propose RACE (Ranking-uncertainty Aware Capacity-Adaptive Customization for Efficient Sequential Recommendation), a framework that dynamically adapts both model structures and input sequences with uncertainty-aware compression ratios. To ensure capacity-adaptive inference, RACE captures the model's intrinsic uncertainty of its predictions and transforms it into a proactive resource allocator. Specifically, we derive a list-wise ranking-uncertainty metric from the distributional stability of top-$K$ predictions, which serves as a measurable proxy for instance complexity. A lightweight predictor is then trained via self-supervision to anticipate this uncertainty from early-layer features. Guided by this signal, RACE performs instance-specific slimming in two dimensions: a hypernetwork-style channel sensitivity extractor identifies parameter redundancy, while importance-aware progressive sequence filtering mitigates data redundancy. Extensive experiments on four popular datasets demonstrate that RACE outperforms state-of-the-art baselines on accuracy, efficiency, and adaptability.
- [389] arXiv:2611.18245 [pdf, html, other]
-
Title: Probabilities of Causation in Nondeterministic Causal ModelsComments: 24 pages, 6 figuresSubjects: Artificial Intelligence (cs.AI)
We study probabilities of causation in nondeterministic causal models, where mechanisms may remain intrinsically stochastic even after all latent variables are fixed. Focusing on binary treatments and outcomes, we develop a generic identification program that represents observational and interventional data as moment constraints on latent stochastic mechanism types and computes minimal/maximal values of counterfactual quantities via constrained optimization. Within this framework, we derive sharp analytic bounds on the probabilities of causation under several assumptions tailored to nondeterministic models. Resulting identification regions are often more informative than those attainable under deterministic alternatives. We illustrate the practical relevance of these results in a case study on personalized treatment decisions and experiments on semi-synthetic datasets. Our results clarify how classical bounds relate to nondeterministic causal semantics, generalizing previous works and providing a unified route from qualitative structural restrictions to computable partial identification regions.
- [390] arXiv:2611.18248 [pdf, html, other]
-
Title: Dancing Tokens on the Vocabulary Network from LLM GenerationComments: 24 pages, 6 figuresSubjects: Artificial Intelligence (cs.AI); Disordered Systems and Neural Networks (cond-mat.dis-nn)
Large language models (LLMs) generate text one token at a time, yet how a particular generation relates to the geometry learned in pre-training remains unclear. The fundamental explainability of LLMs is elusive. Here, we build a similarity network from the input embeddings of six open LLMs and map generation onto it. The token-embedding network is scale-free while heavy-tailed, as are random embeddings with matched covariance, but far more fragmented. Distillation barely changes this topology because input embeddings are largely inherited. The knowledge space is thus deterministic from pre-training. Consecutive generated tokens are rarely graph neighbours and, in most models, no more often than in shuffled order, so generation is not a walk on the predefined network, although the tokens it visits are concentrated within it. Through the logit lens in a layer- wise manner, tokens settle late in depth in most models, and repetitive models remain confined to small domain-specific sub-graphs. These observables could serve as inference-time monitors of degeneration. We thus provide a fresh perspective on understanding LLM generation in the microscopic way from node or link prediction by graph learning.
- [391] arXiv:2611.18251 [pdf, html, other]
-
Title: Lens: Training-Free Semantic Search for Cross-Tile Object and Scene Retrieval in Satellite ImagerySubjects: Computer Vision and Pattern Recognition (cs.CV)
We developed Lens, a training-free system for semantic object and scene search across large satellite mosaics that requires no detector trained for the class of interest. Lens represents each tile with a frozen, self-supervised vision transformer (DINOv3), removes the positional bias in its dense features by projecting them onto the subspace orthogonal to a positional basis estimated once from a constant image, and centers the debiased features on the mean feature computed across the whole area before comparison. Lens forms a query by averaging the centered features of a small neighborhood of patches around a chosen point, or of an entire tile for scene-level search, and matches it against the index by cosine similarity. It stores the centered patch features in a byte-quantized similarity index that supports real-time range search over a mosaic of thousands of patches, and a coarser whole-tile embedding index for scene-level retrieval. A guided upsampling network (AnyUp) runs once per tile to recover pixel-accurate boundaries beyond the native feature grid, and the result is cached so later queries reuse it. We evaluated Lens on the 5,128-tile DFC2020 test set and ran it as an interactive tool on five Sentinel-2 areas spanning agricultural, industrial, and desert terrain; the pipeline is not tied to a specific sensor. On DFC2020 a single click reaches 0.09 to 0.80 pixel F1 across eight land-cover classes at the per-class best threshold, with a mean F1 of 0.39, and it beats a pairwise INSID3 baseline on six of the eight classes at roughly four hundred times lower query cost. Indexing 1.31 million patches takes 169 s on one CPU thread, and a click returns a rendered overlay in 6 to 10 ms.
- [392] arXiv:2611.18252 [pdf, html, other]
-
Title: ParetoM$^3$: Learning on the Pareto Set under Preference Guidance via Min-Max-Min OptimizationComments: Accepted at the 40th Annual Conference on Neural Information Processing Systems (NeurIPS 2026)Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Optimization and Control (math.OC)
Learning across multiple inherently conflicting objectives requires selecting desirable trade-offs from the Pareto set. In this work, we study preference-guided multi-objective optimization, where a predefined preference function is minimized over the weak Pareto set of multiple nonconvex objectives. Leveraging a merit function that vanishes on the weak Pareto set, we propose a penalty-based min-max-min formulation, termed the M$^3$ problem, which unifies preference optimization and the enforcement of weak Pareto optimality in a single objective. We further develop ParetoM$^3$, a single-loop first-order algorithm for computing stationary solutions of the resulting M$^3$ problem. Under a local Kurdyka-Lojasiewicz assumption, we prove that ParetoM$^3$ finds an $\epsilon$-stationary point of the M$^3$ problem that is also approximately weakly Pareto optimal for the original multi-objective problem when the penalty parameter is sufficiently large, with explicit non-asymptotic convergence rates. We also show that this stationarity notion recovers existing optimality criteria, including approximate preference stationarity and approximate Karush-Kuhn-Tucker conditions, under the same assumptions used in prior works. Experiments on a synthetic benchmark, image classification, regression, and math reasoning demonstrate that ParetoM$^3$ reliably attains preferred weak Pareto optimal solutions while achieving strong practical performance. Code is available at this https URL.
- [393] arXiv:2611.18253 [pdf, html, other]
-
Title: Pose-based Neural Clothing for Animated CharactersJulian N. Heidenreich, Vinicius C. Azevedo, Jakob Buhmann, Lento Manickathan, Arnold Moon, Paul Kanyuk, Amit H. Bermano, Jingwei TangComments: 19 pages, 18 figures, 5 tables. Includes appendix. Accepted to the ACM SIGGRAPH / Eurographics Symposium on Computer Animation (SCA 2026); to appear in Computer Graphics ForumSubjects: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV)
Traditional garment simulation for animated characters is a time-consuming bottleneck for artists due to slow simulations and tedious parameter this http URL data-driven methods offer a powerful alternative, capable of learning from a dataset of simulated data to generate high-quality deformations. A lightweight network design allows data-driven systems to achieve a real-time inference speed. This significantly improves the artist workflow, as they can interactively pose a character and see high-quality folds and wrinkles at animation time, thereby allowing to make creative decisions ahead of time. However, existing neural clothing methods have presented appealing results on datasets based on realistic human models (\eg SMPL), but when trained on animated character datasets, they struggle to generalize to unseen poses. In this paper, we aim at designing neural networks and training strategies to generalize better to unseen poses in real-time to enable the interactive posing applications. We focus on a production-oriented setting where the model is trained for a single garment in a quasi-static setup. Technically, we leverage a smooth, expressive latent space by first training a Variational Auto-Encoder (VAE). We then progressively train the VAE encoder to transition from reconstruction to pose-driven prediction. To guide this prediction, we condition the network on both low-dimensional pose parameters and character body surface displacements. Finally, we introduce a novel local gradient histogram loss that mitigates the over-smoothing effect of reconstruction losses. Extensive evaluations on production-level assets and public datasets demonstrate that our method produces sharp, realistic wrinkles, smooth large-scale drape, and temporal coherence in real-time, offering a robust and easily integrable solution for interactive posing workflows.
- [394] arXiv:2611.18255 [pdf, html, other]
-
Title: Linear Modeling of Hybrid Shipboard Power Systems under Pulsed Power Load OperationComments: Accepted and presented at ESARS-ITEC 2026. Copyright 2026 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other usesSubjects: Systems and Control (eess.SY)
A model-based approach is a valuable tool in the power system design process, enabling the integration of component models and the evaluation of dynamic interactions. Simplified models support rapid preliminary assessments by allowing parameters and configurations to be adjusted. This paper investigates the effects of high-bandwidth loads in hybrid AC--DC shipboard power systems. The problem is addressed by developing a linear model aimed at capturing the relevant dynamics over the time scale of interest with reduced computational effort. A LVDC subsystem has been developed with a battery energy storage system that supports the grid during pulsed power load operation. A control strategy has been proposed to satisfy transient requirements. The proposed approach is validated under different pulsed power load conditions using a high-fidelity model developed in DIgSILENT PowerFactory, which preserves nonlinearities and includes converter switching effects and digital control implementation. The validation shows that the model captures the relevant transient behavior with significantly reduced simulation time, while maintaining acceptable accuracy, making it suitable for rapid preliminary design studies.
- [395] arXiv:2611.18256 [pdf, html, other]
-
Title: Deep Bundle AdjustmentComments: 15 pages, 13 figures, 6 tablesSubjects: Computer Vision and Pattern Recognition (cs.CV)
We present \textbf{DeepBA} (\textbf{Deep} \textbf{B}undle \textbf{A}djustment), a graph neural network (GNN) that recovers camera poses, intrinsics, and scene structure from 2D matches across an arbitrary number of views. Rather than solving a non-linear least-squares problem per scene, DeepBA amortizes bundle adjustment into a learned graph network that iteratively refines its estimate from geometric residuals. Where classical SfM computes its updates analytically and needs careful initialization, DeepBA learns them and converges from a cold start. It is trained with scale-invariant supervision (ground-truth rotation, translation direction, and intrinsics, plus reprojection self-consistency), so datasets with arbitrary metric units can be mixed without per-scene normalization. DeepBA is decoupled from the matcher. At 2.04M parameters, it outperforms prior learned SfM from matches on scene reconstruction and on most pose metrics, is competitive with classical SfM at moderate accuracy thresholds and overtakes it in wall time as the view count grows, and approaches monolithic image-to-geometry transformers such as VGGT ($\sim$1.2B parameters) with greater modularity. Project~page: this https URL
- [396] arXiv:2611.18257 [pdf, html, other]
-
Title: NER-D: A cross-domain benchmark for German NERSubjects: Computation and Language (cs.CL)
We present NER-D, a cross-domain benchmark for German NER comprising 13 datasets spanning 9 domains, including General/Politics, Wikipedia, Legal, Medical, Historical, Financial, Agriculture and Archaeology. We examine the performance of both proprietary LLMs and state-of-the-art small open-source models using normal prompting and further examine the performance gains achievable via in-context learning and LoRA fine-tuning. Our results show that without any sort of optimization a proprietary model (GPT-5.4 mini) outperforms the smaller open-source models on 6 out of 13 datasets, but LoRA fine-tuning enables the smaller models to substantially outperform the proprietary baseline. We further analyze the influence of the domain and the tagset used in each task on the performance achieved in the different evaluation settings. Our benchmark will be available as a resource for the development of future German NER methods on both general and specialized domains.\footnote{We make our code, data, and models publicly available upon acceptance of this paper.
- [397] arXiv:2611.18261 [pdf, html, other]
-
Title: Structural Self-Teaching for Compositional Generation in Unified Multimodal ModelsComments: NeurIPS 2026Subjects: Computer Vision and Pattern Recognition (cs.CV)
We propose Structural Self-Teaching (SST), a novel framework that enhances compositional image generation in diffusion-based unified multimodal models (UMMs) by leveraging their internal understanding path as a source of structural supervision. The generation path of UMMs often struggles with dense compositional prompts, leading to missing objects, attribute binding failures, and spatial errors. In contrast, their understanding path localizes the same phrases in images that satisfy these prompts. SST converts this phrase-level grounding into training-time structural signals for the generation path. Specifically, we extract phrase-conditioned support masks from internal attention maps derived from the understanding path and employ them through two mechanisms. Phrase-local contrastive alignment uses these masks as pooling regions and aligns each generated phrase region with the same region of the target image. A structural scaffold builds a phrase-level layout from the same masks and modulates an intermediate spatial state of the generation path during training. The scaffold is phased out during training and removed at inference. SST requires no external structural annotations for training and no additional input at test time. Extensive experiments across multiple UMMs demonstrate that SST improves compositional fidelity over LoRA fine-tuning on the same data, with no inference-time cost. Project page is available at this https URL
- [398] arXiv:2611.18262 [pdf, html, other]
-
Title: Exact-Quota Facility Location: A Transfer Framework and Three Approximation GuaranteesComments: 24 pages. Includes an exact rational certificate and a solver-free verifier as ancillary filesSubjects: Data Structures and Algorithms (cs.DS)
Facility location with outliers asks which facilities to open and which clients to serve, subject to a prescribed service quota. We give a framework for transferring facility-location cost certificates to this exact-quota setting. The framework yields deterministic polynomial-time approximation guarantees of $571/376<1.52$ for metric facility location with outliers, strictly less than $1.853$ for priority facility location with outliers, and $2.16995<2.17$ for squared-metric facility location with outliers. Each algorithm serves the prescribed number of clients and pays the original facility-opening costs.
The framework separates initialization from improvement. A growth process is stopped when it first reaches the quota; deleting excess clients from its last batch leaves a payment deficit supported on a bounded number of facilities. Greedy additions then reoptimize the served client set. A threshold-coverage representation makes the resulting service cost supermodular and allows the usual density analysis to compare different client sets. Finally, enumerating and prepaying a constant number of expensive facilities absorbs the exceptional costs. A transfer theorem states exactly which initialization certificate is sufficient and how many prepaid facilities it requires.
The three applications supply this certificate in different ways. The metric case uses a quota prefix of reassigning JMS; the priority case uses persistent contributions and a cleanup ordered by service level; and the squared-metric case uses a local factor-revealing certificate valid for arbitrary client subsets of the same run. We provide complete finite proofs of the numerical interfaces, including an exact rational certificate for the squared-metric bound. - [399] arXiv:2611.18264 [pdf, html, other]
-
Title: MamaReason: Unfinished Reasoning and Answer Recovery in Clinical Language ModelsThanni Adewuyi, Adepoju Jeremiah Moyondafoluwa, Dorcas Amapu, Dieko Whenu, Sylvia Essessien, Sandra Oko-Oba, Pelumi Onafadeji, Lily Aimas-Love, Samuel Omolaiye, Abiodun AdereniSubjects: Computation and Language (cs.CL)
Answer continuation often returns a diagnosis already stated in unfinished clinical reasoning. We study this behavior in MamaReason, adapted from dense Gemma 4 31B, on MamaBench's 217 paired maternal and child-health presentations. An explicit-commitment extractor identifies a diagnosis in 75 of 94 failed RL first passes and 75 of 93 failed SFT first passes. The continuation matches that diagnosis in 67/75 RL cases and 58/75 SFT cases, using diagnosis strings alone. Reading these commitments without additional generation recovers 30 of the 40 RL pairs restored by continuation under the archived correctness labels. Recurrence controls distinguish routine mentions from repeated answer selection. An offline second-commitment trigger measures the tradeoff between earlier readout and paired accuracy. The resulting decomposition separates diagnostic content already present in a trace, completion of an accepted answer, and paired diagnostic correctness. For clinical language-model evaluation, these are distinct outcomes that become visible when the same reasoning trace is examined through multiple readouts.
- [400] arXiv:2611.18267 [pdf, html, other]
-
Title: TetraCoder: Adaptive Tetrahedral Encoding for Neural FieldsComments: 8 pages, 4 figuresSubjects: Graphics (cs.GR); Computer Vision and Pattern Recognition (cs.CV)
Implicit and explicit representations of geometry are central in computer graphics applications and come with their pros and cons. With surface sampling being faster on explicit representations and spatial queries being faster on implicit representations, finding efficient conversions between them is a highly desirable property. We present TetraCoder, a single resolution hash-free adaptive tetrahedral encoding that links an explicit tetrahedral spatial encoder with an underlying implicit signed distance field (SDF). The resulting hybrid implicit-explicit space is continuous piecewise affine (CPWA) by design, allowing to obtain an exact correspondence between the implicit SDF and the explicit analytic mesh. TetraCoder's design comes with additional perks, in particular: (a) it scales better than multi-resolution encoding methods due to its single resolution optimizable grid, (b) unlike hash-based methods, it is not affected by collisions, (c) it enables fast tetrahedral ray-tracing, geometry and appearance editing, directly through its adaptive tetrahedral scaffold. We demonstrate the versatility of TetraCoder by showing an improvement over state-of-the-art methods on multiple tasks: 2D/3D fitting from images and point clouds, 3D novel view synthesis, 3D reconstruction, and geometry/appearance editing.
- [401] arXiv:2611.18272 [pdf, html, other]
-
Title: From Bytes to Behavior: Request-Level Attribution from Live GPU Hidden StatesComments: 7 pages, 1 figure, 4 tablesSubjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
Existing memory forensics for machine learning systems typically targets static artifacts, such as model weights or prompt logs, rather than the transient computational states of a live inference process. This gap limits an examiner's ability to analyze the specific behavior of a running large language model (LLM) service that cannot be restarted or modified. We present a method for acquiring per-layer residual-stream activations from a live GPU process using a non-modifying debugger, without requiring prior knowledge of the model's architecture parameters or weights. The approach employs a general six-stage process with platform-specific adapters, utilizing either a PyTorch allocator clock or a kernel-launch clock to identify and read tensor allocations from video memory. We validate this acquisition on 33 open-source decoder-only models across nine architecture classes, demonstrating that the recovered bytes are byte-exact and usable for downstream analysis. As a use case, we apply linear probes trained on standard framework activations to the recovered states, successfully distinguishing refusal from compliance with an AUROC of 0.88 on held-out prompts. This work establishes that live, observational access to intermediate hidden states is feasible on named inference platforms, providing a new source of evidence for examining the operational behavior of deployed AI systems.
- [402] arXiv:2611.18274 [pdf, html, other]
-
Title: Right Token, Right Time: From Per-Token Gating to Trajectory-Aware Activation SteeringSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Activation steering via sparse autoencoders (SAEs) enables behavioral control of large language models without task-specific fine-tuning. Yet, standard methods apply the steering signal at every generated token, displacing representations from the distribution the model was trained on and eroding generation quality. We show that \emph{when} along a generation the intervention is applied is a decisive axis of control. We cast steering as a per-token allocation problem under a fixed budget of steerable tokens, so that methods differ only in which tokens they select, making allocation the sole variable. Within this framework we introduce two gates. The first, Contrastive Feature Projection (CFP), is memoryless and selects the tokens least aligned with the target behavior through a sparse, interpretable feature projection. The second, Closed-Loop Adaptive Bernoulli Steering (CLABS), is trajectory aware. It treats allocation as closed-loop feedback control, tracking the alignment a generation has accumulated so far and spending budget where the trajectory falls behind. On a linear steering-response model we derive an explicit gain region that guarantees closed-loop stability of CLABS and show that its integral term reduces the steady-state tracking offset. Empirically, CFP and CLABS achieve the largest behavioral shift among budgeted gates at matched budgets, with CLABS strongest under the tight budgets where allocation matters most, while best preserving fluency.
- [403] arXiv:2611.18275 [pdf, html, other]
-
Title: BPU-SpikeMM:Exact Matrix Multiplication via Spatio-Temporal Spike Encoding on Real Neuromorphic HardwareSubjects: Distributed, Parallel, and Cluster Computing (cs.DC)
Matrix multiplication is the dominant computational kernel in deep learning, yet its growing scale strains conventional von Neumann hardware in energy efficiency and bandwidth. Neuromorphic architectures offer a promising alternative, but mathematical equivalence between matrix multiplication and spiking computation does not guarantee direct execution on real neuromorphic hardware, as the specific hardware paradigm may still require additional data recovery, bit-weight reconstruction, or state readout after neuromorphic execution. This paper presents BPU-SpikeMM (Brain Processing Unit-based Spike Matrix Multiplication), an exact spatio-temporal spike encoding framework for matrix multiplication on real neuromorphic hardware. We prove the exact equivalence of spike-encoded matrix multiplication under the BPU execution paradigm, distinguish mathematical equivalence from native direct-readout equivalence on BPU hardware, and build a spatio-temporal offloading pipeline supporting multiple encoding schemes. Furthermore, we extend the framework to a multi-dimensional encoding design space and evaluate the applicability of different encodings on real BPU hardware in terms of correctness, temporal resource cost, spatial resource cost, and event-driven metrics. The results confirm native direct-readout equivalence across all evaluated encoding schemes and show that their applicability to BPU execution depends on joint trade-offs among temporal cost, spatial resource consumption, and event activity.
- [404] arXiv:2611.18279 [pdf, html, other]
-
Title: GT-FENG: General Theory of Factual Evaluation and Normative Grounding in Language Models *Author(s)Comments: 7 pages, 4 tables. Code: this https URLSubjects: Computation and Language (cs.CL)
Whether a language model's answer is correct often depends on context: the same value can be right under one jurisdiction, date, or legal definition and wrong under another. We introduce GT-FENG (General Theory of Factual Evaluation and Normative Grounding), the first general framework to unify contextual/temporal correctness and rule-based error attribution across law, policy, and other rule-governed domains. GT-FENG checks answers against the applicable rules and identifies documented rules that match incorrect answers. It shows that when multiple rules give the same value, the answer alone cannot tell which rule the model used. In Medicaid income-limit answers from two models, documenting the asked state's own records reduces the number of wrong answers that match only other states' values by 68-78%, and more than half of stable wrong answers then match two or more documented sources. GT-FENG provides a foundation for evaluating language models in rule-governed domains and for future benchmarks and evaluation toolkits. We release GT-FENG as an open-source Python package, gtfeng.
- [405] arXiv:2611.18284 [pdf, html, other]
-
Title: Clean Tickets, Wrong Orders: Catching Silent Errors with a Language Model in Your Browser TabSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
A small language model that turns a typed order into structure can fail in two ways. It can print something broken, which any system rejects. Or it can print a clean, well formed order that nobody placed. Our companion work calls the second kind a silent error, and it is the dangerous one. We present a web demo in which a fine-tuned 0.8B parameter model runs entirely inside the visitor's browser tab, through this http URL compiled to WebAssembly. No order text leaves the device. The page parses orders as the visitor types, scores every ticket with the PIZZA benchmark's own matcher ported to TypeScript, refuses anything off the menu with "?" so the model can face the open web, and lets visitors hunt for silent errors or try to cause one. Run in the browser on all 1,357 PIZZA test orders, the deployed model writes the same output as the server evaluation for 99.0% of them and reaches the same exact match (0.882); we also report its latency on 3 consumer laptops. Demo: this https URL
- [406] arXiv:2611.18293 [pdf, html, other]
-
Title: PrivacyPR: A Dataset of Real-World Privacy Fixes for Benchmarking Privacy-Aware Code ReviewComments: 4 pagesSubjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
Coding agents increasingly review code, yet whether they catch the privacy defects that laws such as the GDPR require software to avoid remains largely untested, partly because few datasets connect real privacy defects to their accepted fixes and review context. We present PrivacyPR, a dataset of 4,131 merged GitHub pull requests (PRs) from 1,469 repositories, created between June 2018 and January 2024. Each PR includes its commits, diffs, review discussion, and LLM-generated labels for privacy relevance, fix type, and GDPR principles. PrivacyPR provides 1,169 high-confidence privacy fixes and 1,643 non-fix PRs, together with Verified, a manually checked split of 250 privacy fixes and 250 non-fix PRs, and Small, a 100 + 100 subset for evaluation. On Verified, both annotators endorsed 72% of the LLM's GDPR-principle assignments. Beyond supporting empirical study of how privacy defects arise and are fixed, PrivacyPR provides a benchmark for privacy-aware code review on real, developer-accepted fixes. Its evaluation harness reverts each fix to reintroduce the defect. On Small, two coding agents detect only 42-51% of defects under a generic review prompt, but 88-91% when explicitly asked to check privacy, while flagging 11-15% of non-fix PRs. These results show that privacy defects can be detected by coding agents, but that detection depends strongly on review intent and comes with false alarms. Data, code, reviews, judge verdicts, and an extensible evaluation harness are publicly available at this https URL.
- [407] arXiv:2611.18297 [pdf, other]
-
Title: When evaluation becomes intervention in clinical AIAnthony Abou Mrad, Aviona Conti, Samuel Schmidgall, Jayanth Pratap, Ruhika Iyer, Joseph Kim, King Shi, Yann Sakref, Luis M. Ahumada, Lisa Knowlton, Andrew M. Cameron, Khyzer Aziz, Jeffrey K. JoplingComments: 12 pages, 3 figures, 1 box. Submitted to npj Digital MedicineSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY)
AI judges, models that grade another system's output against a rubric, extend clinical review to the scale at which clinical AI writes. When teams use judge assessments to select a prompt or train a model, the judge becomes part of the intervention. Selection can then favor errors the judge overlooks. We describe what a judge's report must contain and the independent clinical test each judge-driven change needs.
- [408] arXiv:2611.18298 [pdf, html, other]
-
Title: Objectives as Policies: Vision-Language Model Cost Programs for Contact-Rich ManipulationComments: 8 pages, 3 figures. Project website: this https URLSubjects: Robotics (cs.RO)
Vision-language models (VLMs) have shown strong semantic and spatial reasoning capabilities in 3D environments. Recently, connecting VLMs with model-predictive controllers (MPC) has emerged as a paradigm for combining strong semantic reasoning with physics-based planning. However, connecting these two for contact-rich robotic manipulation requires a task representation that bridges the two and captures how physical interactions can be reasoned by VLMs. In this work, we introduce Objectives as Policies (OaP), a framework that connects a pretrained VLM to a shared sampling-based MPC through an editable objective program. Each program specifies a sequence of stages for completing a task, combining running and terminal costs from a cost library with stage completion conditions. Given a natural-language task instruction and scene observation, we build a simulation-based digital twin through real-to-sim, and the VLM generates an objective program that is executed by the MPC in the simulation. The program is iteratively updated by the VLM based on MPC rollouts, before being deployed in the real world. We evaluate our framework in both simulation and the real world on four tabletop manipulation tasks, and show that it outperforms strong baselines. Additional videos and examples are available on the project website: this https URL
- [409] arXiv:2611.18305 [pdf, html, other]
-
Title: PROOF: Perceptive Robustness via Out-of-Distribution Fallback for Legged LocomotionComments: CoRL 2026Subjects: Robotics (cs.RO)
Vision-based perceptive locomotion has enabled legged robots to achieve agile and terrain-aware locomotion across complex environments. However, the reliability of such systems heavily depends on visual observations, which are inherently susceptible to occlusion, reflective surfaces, and illumination variations, often resulting in catastrophic locomotion failures under degraded perception conditions. In this paper, we present PROOF (Perceptive Robustness via Out-of-Distribution Fallback for Legged Locomotion), a perceptive locomotion framework that enables robust locomotion under severe visual disturbances. Our method adopts a two-stage teacher-student learning paradigm that explicitly models perception reliability: the teacher learns both perception-driven and proprioceptive fallback behaviors, while the student predicts exteroceptive features with uncertainty and suppresses them under out-of-distribution inputs. We evaluate our method on a Unitree Go2 quadruped equipped with a RealSense D435 depth camera in real-world settings. Under normal visual conditions, our method enables the robot to successfully perform high-dynamic parkour maneuvers, including traversing large gaps, elevated platforms, and narrow beam terrains. Under severe visual degradation, such as heavy occlusion and strong illumination changes, the robot continues to maintain stable and robust locomotion performance.
- [410] arXiv:2611.18309 [pdf, other]
-
Title: Kronecker Approximations of Covariance Matrices for Solving Inverse ProblemsSubjects: Numerical Analysis (math.NA)
Inverse problems arise in many applications where the task is to recover unknown parameters from indirect or noisy observations. In contrast to forward problems, where outputs are directly computed from known inputs, inverse problems are designed to infer the underlying variables from the observed effect. Inverse problems are often ill-posed, so regularization is needed to compute reasonable solutions. One form of regularization incorporates prior knowledge about the unknown parameters in the form of prior distributions. However, in many settings, large prior covariance matrices pose computational bottlenecks for solving inverse problems. To address this, we study approximations of large covariance matrices via structured Kronecker-Toeplitz products and demonstrate how these approximations can be used to solve inverse problems. We exploit the Block-Toeplitz-Toeplitz-Block structure induced by the covariance kernel to construct compact approximations that are orders of magnitude cheaper to store and invert. These approximations are deployed as prior covariance surrogates in Bayesian inverse problems solved via generalized hybrid iterative methods. We demonstrate that the Kronecker-Toeplitz approximation attains lower relative approximation error than existing matrix approximations, exhibits rapid error decay with rank, and enables efficient matrix-vector products. Applications to 2D seismic tomography reconstruction problems illustrate the practical utility of these methods.
- [411] arXiv:2611.18312 [pdf, other]
-
Title: On the Same Track: Privacy Leaks in Electric Rail Transport via Magnetic FieldsSubjects: Cryptography and Security (cs.CR)
Inertial sensors of smartphones are a well-known privacy risk, as they allow inferring sensitive user movements and activity without explicit permissions. Among these sensors, the magnetometer has received little attention, as it is assumed to measure mainly the Earth's magnetic field. In this paper, we show that electric trains generate distinctive magnetic fields that can be captured by a smartphone magnetometer. Apps deployed on the device can access these signals without user permission, enabling two novel attacks on passenger privacy: First, we are able to determine whether the user is riding an electric train (transport mode detection). Second, we can infer whether other victims are aboard the same train due to shared patterns in the magnetic field (colocation prediction). In an empirical evaluation across 217 train rides, we demonstrate that both attacks are effective, confirming a real-world threat. We discuss countermeasures and argue that magnetometer access should require explicit user permission in mobile operating systems.
- [412] arXiv:2611.18313 [pdf, html, other]
-
Title: Hierarchical Interactions in Mobility Networks: A Model-Free Feedback ApproachComments: 12 pagesSubjects: Systems and Control (eess.SY); Optimization and Control (math.OC)
Urban mobility systems are large-scale networked systems whose performance emerges from hierarchically coupled decisions of public authorities, competing mobility operators, and travelers. Existing approaches to policy design either collapse one of these levels or assume that the authority can model and differentiate through the equilibrium responses of operators and travelers. In this paper, we model policy design as a tri-level game on a multimodal network, in which a municipality sets taxes, licenses, and integration subsidies, a public transport and a taxi operator compete on prices and service supply, and heterogeneous travelers reach a stochastic user equilibrium. We prove that the traveler equilibrium is unique and continuously differentiable in the operator strategies, that the variational inequality characterizing operator equilibria admits a solution for every policy, and that the resulting tri-level game admits a Stackelberg equilibrium. Building on this structure, we design a model-free hierarchical controller in which the municipality and the operators update their decisions from zeroth-order evaluations of observed equilibrium outcomes. On the Zurich multimodal network, operator equilibria computed from different initializations differ by a median of 0.3% in the municipal objective. On a grid of 3,125 policies, only those that subsidize transfers between public transport and taxis benefit both operators, and 24 of them also lower travel costs without degrading the municipal objective. Finally, the controller matches the best policies of three black-box optimizers and terminates at the most consistent policies across runs.
- [413] arXiv:2611.18315 [pdf, html, other]
-
Title: Experts Across Schemas: Agent-Guided Mixture-of-Relational-Experts for Cross-Database GeneralizationSubjects: Artificial Intelligence (cs.AI)
Web platforms record user and item activity in relational databases whose schemas differ from platform to platform, which makes it difficult to transfer predictive models between them. Relational foundation models address this through cross-database pretraining, yet recent improvements have relied on increasingly large pretraining corpora. We ask whether a better architecture, together with features specific to the target database, can deliver comparable gains from far less pretraining data. We introduce MoREGT, a mixture-of-relational-experts graph transformer that represents both rows and cells as tokens, adds the log-ages of rows and the names of foreign-key relationships to attention, and routes cells and rows to experts by column metadata rather than by values. A label-blind LLM agent complements the model: it reads the target database's schema and column statistics and adds validated single-row derived columns, without access to labels or model scores. On 21 RelBench tasks with leave-one-database-out pretraining on six databases, the MoREGT backbone improves on the Relational Transformer pretrained on the same databases by 2.25 points of mean AUROC and 5.83 points of mean nMAE. The agent's columns raise mean AUROC from 78.45 to 79.63 and lower mean nMAE from 27.91 to 26.98, the highest means on both metrics among the compared methods, including RT-J, which pretrains on about 400 times as many databases.
- [414] arXiv:2611.18319 [pdf, html, other]
-
Title: Learning Word-Level Explanations of Text Reuse in Latin Literature with Little AnnotationSubjects: Computation and Language (cs.CL)
Detecting an intertextual reference is only the first step; philologists also need to know how a later author reworked the source. We explain a reuse pair word by word: each word of the reusing passage is first aligned to its source word, then labeled with the operation that transformed it, such as a copy, a change of inflection, or a substitution. Such word-level operations are expensive to annotate, and only a few corpora annotated at this level exist for text reuse in Latin literature. We describe a procedure that needs only a limited number of annotations: synthetic pairs, built by applying known operations to Latin text, train a first model, and active learning chooses the pairs an annotator corrects. Among several architectures, a model that links and labels together performs best: a pointer network that, for every reuse word, chooses a source word and names the operation in the same step. We show that with active learning, 200 pairs chosen by uncertainty are enough for the model to match one trained on every annotated pair of the 1,490 references of the Loci Similes benchmark. We release our code, word-level predictions for all references, and an interactive demo.
- [415] arXiv:2611.18321 [pdf, html, other]
-
Title: Causal AI-Driven Secure mmWave Beam Alignment with Defensive Variational AutoencodersSubjects: Artificial Intelligence (cs.AI)
Efficient and secure beam alignment is essential for mmWave communication systems. However, existing deep learning (DL)-based beam alignment methods often neglect causal relationships between inputs and outputs, resulting in limited interpretability, poor generalization, and excessive beam sweeping. Moreover, these DL solutions are vulnerable to adversarial attacks, where small perturbations in received power measurements can mislead beam classifiers. To overcome these issues, we propose a causally-aware DL framework integrating causal discovery into beam alignment. A two-stage causal beam selection algorithm learns a Bayesian graph capturing dependencies between received signal strength indicators (RSSI) as inputs and the optimal beam index as the target variable, then uses it to select a minimal set of relevant input features for the DL classifier. To mitigate adversarial attacks, we introduce a defensive variational autoencoder (DVAE) that maps potentially perturbed inputs back to their clean versions. End-to-end fine-tuning of the downstream beam classifier on the DVAE-reconstructed RSSI representations further boosts prediction accuracy. Simulation results show that causal beam selection achieves accuracy comparable to conventional methods while providing an \(\mathbf{829\times}\) lower per-sample computational cost than SHAP (Shapley Additive Explanations), a widely used explainable AI (XAI)-based feature selection method. It further reduces beam-sweeping overhead by \(\mathbf{59.4\%}\) and \(\mathbf{89.8\%}\), compared to conventional DL-based beam selection and exhaustive search, respectively. The DVAE-based defense effectively denoises adversarial inputs, sustaining 85\% accuracy across different perturbation bounds, retains baseline performance on unperturbed RSSI inputs and surpasses defensive distillation defense scheme by 15--20\% on average under different attack scenarios.
- [416] arXiv:2611.18322 [pdf, html, other]
-
Title: Power Management of Electrically Driven Centrifugal Compressors for Flexible OperationSubjects: Systems and Control (eess.SY)
We present a control strategy for power management of a variable-speed 15MW (rated power) centrifugal compressor, and investigate how quickly it can follow a reduced electrical power reference without entering surge. We use a Greitzer-Gravdahl type model scaled to industrial size with rate-limited valve actuators as a simulation environment and propose an advanced regulatory control structure using feedback control loops with dynamically scaled feedforward mechanisms and a structure of min/max selectors. In simulated experiments, a 6MW power reduction from varying initial operating points on a 12MW iso-power line is delivered at rates of up to 10MW/s, this http URL 90% in 0.54s, without opening the recycle valve. Depending on the initial operating point, a large power reduction can be delivered at once or with a restricted ramp rate. The stroking rate of the discharge side valve is identified as the limiting factor for how fast the power can be reduced in the studied configuration. A limited sensitivity study on the impact of rotor inertia for power shedding performance is also provided.
- [417] arXiv:2611.18324 [pdf, html, other]
-
Title: Unlocking Cross-Attention in Label-Scarce Regimes: A Weakly Supervised Approach for Video Semantic SegmentationComments: Accepted at WACV 2027. 25 pages (10 main + 15 supplementary), 9 figures, 7 tablesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Weakly Supervised Video Semantic Segmentation (WS-VSS) aims to learn temporally consistent semantic object masks in videos using only video-level labels. While fully supervised models successfully learn temporal relationships by means of cross-attention, training these temporal weights under weak supervision is unstable and even pushes the model below its single-frame baseline. Therefore, existing WS-VSS methods exploit temporal consistency by relying on external priors, such as optical flow. Unfortunately these cues are non-semantic, computationally expensive and noisy, degrading the supervisory signal. Instead of supervising attention weights directly or using external priors, we propose MemCAM, a WS-VSS transformer framework that learns temporal cross-attention directly through feature-space self-supervision. MemCAM introduces a CAM Propagation Algorithm that synthesizes a supervisory signal by propagating past CAMs according to patch-level feature similarities in the latent space. Specifically, we constrain their induced CAMs through a consistency objective between predicted and propagated maps, stabilizing temporal attention learning. Experiments on the YouTube-Objects v2.2 and WS$^2$ datasets demonstrate that MemCAM consistently outperforms the temporal localization and segmentation performance of WS-VSS alternatives, showing that cross-attention can be effectively trained in label-scarce video settings through internal self-supervision.
- [418] arXiv:2611.18333 [pdf, other]
-
Title: Emulating physics-based XCO2 retrievals across OCO-2 viewing modes with compact neural networksSubjects: Machine Learning (cs.LG)
Operational full-physics retrieval of the column-averaged dry-air mole fraction of carbon dioxide from satellite spectra is computationally demanding. We use singular value decomposition to compress spectra from the Orbiting Carbon Observatory-2 separately by band. Compact multilayer perceptrons use the resulting spectral scores, prior information, and viewing geometry to predict the difference between the full-physics retrieval and its operational prior. Adding the predicted difference to the prior yields an estimate of the retrieved mole fraction. The models were trained on 2015-2019 observations and validated on 2020 observations. In 2021-2025 tests, the mean annual root-mean-square errors relative to the Atmospheric Carbon Dioxide Observations from Space retrieval product were 0.89, 1.35, and 1.43 parts per million for ocean glint, land glint, and nadir observations, respectively. On East Asian nadir observations screened using the published criteria for the spectral Transformer, our model yielded a lower mean annual error than reported for that model in 2020-2022. These results show that compact neural networks using compressed spectra can efficiently emulate the full-physics retrieval across multiple viewing modes and years.
- [419] arXiv:2611.18338 [pdf, html, other]
-
Title: MEM-SQL: Memory-Empowered Text-to-SQLSubjects: Databases (cs.DB)
Text-to-SQL systems translate natural language questions into SQL queries over increasingly complex databases. However, widely used benchmarks evaluate questions independently, overlooking settings where earlier queries over the same database can supply joins or calculations for a new question. We introduce HARE (History-AwaRe Text-to-SQL BEnchmark) to measure the accuracy and cost of adapting historical SQL. HARE comprises 1,600 requests derived from a fixed pool of 823 verified queries across four datasets. Annotations record which components each request preserves and which edits it requires, distinguishing whole-query adaptation from partial reuse. Our method, MEM-SQL (Memory-EMpowered Text-to-SQL), stores whole-query templates, reusable SQL blocks, and mappings from domain concepts to schema elements. MEM-SQL uses a layer only if every requirement assigned to it has a match above that layer's threshold. With On-demand Schema Linking (OSL), MEM-SQL consults the schema only when no layer qualifies or the generator judges retrieved evidence insufficient. Across six language models on HARE, mean execution accuracy rises from 37.1% for Direct Generation without history to 50.0% for MEM-SQL, with 76.6% fewer input tokens and 68.8% lower latency on average. Further analysis shows that requests reusing several verified queries remain a bottleneck, with accuracy falling from 56.7% to 29.7%, most sharply when reused components must be placed in a new query structure or combined with a shared join. Applying OSL to three other memory designs consistently reduces generation cost, while its accuracy benefit depends on the memory design. Verified history can thus make generation both more accurate and cheaper, but its value depends on the evidence retained and when the system consults the schema.
- [420] arXiv:2611.18339 [pdf, html, other]
-
Title: Control Invariance on Manifolds through Tangent Bundle SplittingSubjects: Systems and Control (eess.SY)
Control invariance for interconnected non-linear systems on manifolds requires conditions under which subsystem safety constraints can be combined to form an invariant set. This paper develops local product type control invariant sets on manifolds using tangent bundle splitting, adapted product coordinates, and subsystem tangency conditions satisfied simultaneously by one admissible input at each state. The resulting conditions establish control invariance of product sets without requiring independent subsystem dynamics, with separate control choices possible when inputs are independently constrained and each acts only on its subsystem. This provides a way to verify subsystem safety without discarding dynamic coupling and is illustrated by a torus example in which adapted coordinates give a curved constraint region a product representation.
- [421] arXiv:2611.18340 [pdf, html, other]
-
Title: DIRT: Depth-completed Image-based Rendering using Two-band Blending for Live Novel-view SynthesisComments: 8 pages main paper + 2 pages of references + 9 pages supplementary materialSubjects: Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR)
State-of-the-art novel-view synthesis (NVS) methods commonly require per-scene optimization before rendering, limiting their applicability in live settings. Live or ad-hoc NVS, however, demands low-latency rendering during capture and strong generalization to new scenes. Some previous work tried to address this task by adapting training- and optimization-based methods, trading quality for low latency. We introduce a live NVS approach following traditional image-based rendering that operates on a posed set of depth sensor observations and images. Our method leverages explicit, sparse point-based geometry obtained from active sensors, resulting in a sparsely filled depth buffer that is completed in the (novel) target view's screen-space. Using this proxy, a posed set of images is projected into the target frame before fusing the resulting observations through a frequency-aware two-band blending strategy. The pipeline is entirely non-neural and optimization-free, requiring no training data or scene-specific optimization. As a result, our approach reduces the delay between capture and novel-view rendering and does not rely on favorable network generalization. We validate the method's applicability to heterogeneous sensor setups - including handheld RGB-D cameras, automotive LiDAR, and thermal cameras - as well as challenging scenarios, such as low light captures. Our experiments demonstrate that the approach outperforms prior methods in live reconstruction scenarios, while maintaining competitive performance in conventional NVS benchmarks.
- [422] arXiv:2611.18346 [pdf, html, other]
-
Title: VLAN: Vision-Language Accessible NavigationComments: Accepted to NeurIPS 2026 Evaluations and Datasets Track. 25 pages, 17 tables, 6 figures. Code: this https URLSubjects: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
Accessible mobility is user-conditioned rather than scene-intrinsic, as walkable paths for one embodiment are inaccessible to another. Yet, existing benchmarks evaluate traversability under a fixed and limited embodiment. We present Vision-Language Accessible Navigation (VLAN), containing 1,000 scenes, 20K navigation trajectories, and 160K keyframes across diverse lighting, environmental conditions, scene semantics, and path geometries. For each trajectory, VLAN provides embodiment-specific actions and accessibility features for legged and wheeled agents, including humanoid and quadruped robots and people with visual or mobility impairments, while everyday obstacles and accessibility labels are categorized as Passable, Bypass, HighRisk, and Blocked. Unlike goal-only evaluation, our benchmark measures physical safety and alignment with user-specific constraints via goal-reaching, collision-aware, and accessibility-aligned metrics. We benchmark modern vision-language models and find they often fail to consistently adapt obstacle relevance, path selection, and safety criteria across user profiles, suggesting accessibility-aware navigation requires dedicated datasets, tasks, and metrics beyond conventional embodied navigation evaluation. All data and code open-source at this https URL.
- [423] arXiv:2611.18350 [pdf, html, other]
-
Title: MaskSCD: Bitemporal Mask Classification for Semantic Change Detection in Remote Sensing ImagesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Semantic change detection (SCD) identifies where the land surface has changed and which land-cover transition has taken place. Most deep SCD networks predict the change map and the land-cover map of each date with separate per-pixel decoders. These outputs need not agree, so a pixel can be labeled as changed although it has the same class at both dates, and per-pixel decisions fragment objects. We present MaskSCD, which formulates SCD as the mask classification of a bitemporal label-pair image. Each bitemporal segment, a region of constant land-cover class pair, is predicted as one mask shared by both dates together with one class per date, and the change map is derived from the predicted class pair. This formulation keeps the change and semantic maps consistent and the label space linear in the number of classes. A bitemporal masked-attention transformer decoder implements it by reading each segment in both images and relating the two readings, and a bitemporal fusion pyramid provides per-date and change-aware features. MaskSCD is evaluated on SECOND, Landsat-SCD and Hi-UCD, which span 0.1 to 30 m resolution and four to nine land-cover classes, against fourteen recent SCD models re-trained under one protocol. MaskSCD achieves the best separated kappa, semantic change $F_1$ and transition $F_1$ on all three datasets and exceeds the strongest competitor by about five points in mean separated kappa. It produces no contradictory change pixels, which account for nearly 30% of the predicted change of some multi-task networks. The code is available at this https URL.
- [424] arXiv:2611.18355 [pdf, html, other]
-
Title: Information-Theoretic Threshold Designated-Verifier Digital Signatures for Resource-Constrained DevicesSubjects: Cryptography and Security (cs.CR)
Digital signatures are synonymous with data integrity and authenticity. Among applications such as e-commerce, software distribution, and instruction processing they play a vital role in the secure communication of data. Within this context and in preparation for quantum adversarial tactics we present three new information-theoretic threshold designated verifier digital signature schemes and prove them secure against \textit{any} adversary under the standard security criterion. The $n$-of-$n$ multi-signature, $t$-of-$n$ threshold with dealer, and $t$-of-$n$ threshold without dealer designated verifier signature schemes can be implemented with simple mathematical constructs and we demonstrate the efficiency of these with evaluation of signature computation ($\lessapprox$ 80\,ms) and verification times ($\lessapprox$ 150\,ms) on the resource constrained platform of Raspberry Pi 4. We find our scheme also has equivalent or improved (2$\times$) performance than a comparable state-of-the-art post-quantum threshold digital signature scheme on both resource constrained devices and the more capable consumer-grade 2025 M5 MacBook Pro.
- [425] arXiv:2611.18358 [pdf, html, other]
-
Title: Quadratic APN Functions of $\mathbb{F}_2^{10}$ with a Self-Equivalence of Prime OrderSubjects: Information Theory (cs.IT); Cryptography and Security (cs.CR); Combinatorics (math.CO)
Almost perfect nonlinear (APN) functions of $\mathbb{F}_2^n$ have the lowest possible differential uniformity. The only known APN permutation in even dimension, in dimension six, is equivalent to a quadratic APN function with a nontrivial linear self-equivalence; no APN permutation is known for even $n \geq 8$. In a published collection of about 3.8 million quadratic APN classes of dimension eight, at most 0.34% have a nontrivial linear self-equivalence. In dimension ten, the 25 literature classes considered here all have one. We study all quadratic APN functions of $\mathbb{F}_2^{10}$ with such a self-equivalence, which usually confines them to a much smaller linear space. Up to conjugacy and power similarity, self-equivalences of prime order have 299 types. We settle 298 of them. Elementary lemmas exclude 280 types, and exhaustive searches confirm this for 242 of them. The other 18 types are settled by exhaustive searches with exact equivalence certificates. Up to extended affine equivalence, the 298 types contain 29 classes of quadratic APN functions, four of them new. This also classifies the quadratic APN functions with coefficients in a subfield. Exactly 24 classes contain one with coefficients in $\mathbb{F}_4$, and all 3 that contain one with coefficients in $\mathbb{F}_{32}$ are among them. The open type consists of the quadratic canonical triplicate functions. They correspond to maximum additive Hermitian rank-distance codes. Three different searches find only the invariants of 16 classes, all among the 29. We conjecture there are no others. No quadratic APN function in the 298 settled types is Carlet-Charpin-Zinoviev equivalent to a permutation.
- [426] arXiv:2611.18359 [pdf, html, other]
-
Title: Envisioning Agentic Robotics: A New Era of Robotic SystemsGrid Robotics: Qingfeng Lan, Duo Li, Siyao Li, Yiheng Lyu, Shijian Lu, Zhenghao Qi, Shizun Wang, Xinchao Wang, Ming Yang, Zhiqin Yang, Zuhao Yang, Tao YuComments: Project website: this https URLSubjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
The release of GPT-6 Astra marks the emergence of a new paradigm in robotics: agentic robotics. Within weeks, frontier models with agentic harnesses matched or exceeded task-specific policies on many embodied benchmarks. Instead of training a dedicated policy for each family of tasks, an agentic robotic system applies a foundation model to open-world embodied tasks. It operates in a closed loop: observing the environment, interpreting user intent, reasoning about the task, invoking skills and tools, verifying outcomes, and adapting its strategy. Despite these advances, the community has yet to converge on a shared definition, a reliable evaluation framework, or a research roadmap. We argue that Embodied System 2, realized by a foundation model together with an embodied harness, forms the decision-making core of agentic robotics. Its central challenge is to translate general-purpose reasoning into effective and verifiable actions under physical constraints such as limited reaction time, mechanical limits, unpredictable environment changes, and irreversible consequences. Motivated by this perspective, we propose an evaluation framework that measures both model capabilities and closed-loop system performance. Furthermore, we identify the principles that carry over from digital to physical agents and the challenges that must be addressed to achieve reliable physical autonomy. On top of the Embodied System 2 and the evaluation framework, we envision a research roadmap toward sustained autonomy: the ability to perform work safely and reliably with progressively broader responsibility, from executing individual primitives to fulfilling standing roles.
- [427] arXiv:2611.18363 [pdf, html, other]
-
Title: Adaptable Watermarking: Transforming Image Watermarking Models via Lightweight AdaptersComments: Accepted to WACV 2027Subjects: Computer Vision and Pattern Recognition (cs.CV); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
We explore the adaptability of neural image watermarking models; the extent to which one watermarking method can be transformed into another through lightweight adaptation. We introduce a framework that enables encoder-decoder watermarking architectures to become modular and upgradable, supporting multiple watermarking styles without re-training full models. Adaptation is performed by attaching a compact parameter-efficient adapter and fine-tuning under a multi-domain residual alignment loss (spatial, gradient, and frequency) while preserving bit accuracy via a secret-loss term. We evaluate this approach across a suite of recent watermarking models, including TrustMark, InvisMark, RoSteALS, HiDDeN, and WAM, demonstrating that adapted encoders closely reproduce target residuals and that adapted decoders recover messages with near-optimal accuracy under standard augmentations and distortions. We show that cross-model adaptation, with applications in interoperable and upgradable watermarking, can be achieved with as few as 1-2M trainable parameters per component (e.g., the 2M decoder adapter used in our multi-decoder inference pipeline).
- [428] arXiv:2611.18364 [pdf, html, other]
-
Title: Reinforcement Learning for Null-Space Exploitation in Pseudo-Inverse Control AllocationSubjects: Systems and Control (eess.SY)
Pseudo-inverse methods are widely used to solve control allocation problems for overactuated aircraft with redundant control effectors. These methods reliably satisfy the primary command tracking objective, while the remaining null-space degrees of freedom are in principle available for secondary objectives such as drag reduction or actuator wear balancing. However, exploiting this freedom in practice requires hand-tuned weighting matrices or a priori preferred deflections that are difficult to specify across a wide flight envelope with nonlinear, state-dependent dynamics. This paper presents a hybrid architecture in which a reinforcement learning agent operates exclusively within the null space of the control effectiveness matrix, learning secondary objective strategies directly from interaction with the flight dynamics. The null-space projector guarantees by construction that primary command tracking remains invariant regardless of the agent output, providing a principled separation between certified primary control and adaptive secondary optimization. The approach is demonstrated on a nonlinear six-degree-of-freedom distributed electric propulsion aircraft model with aero-propulsive effects across diverse mission scenarios, achieving measurable drag reduction compared to the baseline allocation while preserving tracking integrity and respecting actuator constraints. The proposed architecture forms a foundation for future extensions toward multi-objective, mission-conditioned policies.
- [429] arXiv:2611.18365 [pdf, html, other]
-
Title: Heavy Metal Acrobatics: Learning Dynamic Control for Skid-Steer LoadersPierre-Jean Lorentz, Pascal Egli, Edo Jelavic, Cliff Li, Ramitha Sundar, Harikrishnan Suresh, Paul Montgomery, Lee Redden, Matthias Müller, Ozan Sener, Vladlen Koltun, Marco HutterJournal-ref: P.-J. Lorentz, P. Egli, E. Jelavic, C. Li, R. Sundar, H. Suresh, P. Montgomery, L. Redden, M. M\"uller, O. Sener, V. Koltun, M. Hutter, Heavy metal acrobatics: Learning dynamic control for skid-steer loaders. Sci. Robot. 11, eaef0894 (2026)Subjects: Robotics (cs.RO)
Automation of heavy machinery offers the opportunity to improve productivity in mechanized industries such as construction, mining, and material handling. Yet, automation of such machines is challenged by their complex dynamics, in part because of hydraulic actuation that involves fluids, flexible elements, and nonlinear delays. A concomitant challenge is the size and weight of these machines, which can make experiments on the physical systems exceedingly dangerous. Here, we show that \ac{RL} in simulation, followed by transfer of the learned policies from simulation to reality, can successfully automate acrobatic maneuvers that push a heavy construction machine to the limits of its actuation and have only previously been accomplished by a small number of elite human operators in official world-record performances. We trained policies that perform autonomous acrobatic maneuvers on a John Deere 332G, an operational 5-tonne skid-steer loader. The policies can autonomously swing the machine up onto two wheels and, while balancing on two wheels, drive over rough terrain, trace out specified trajectories, spin around, and swing the bucket back and forth before landing it safely back onto four wheels. This work demonstrates that \ac{RL} and sim-to-real transfer enable dynamic autonomous control of heavy machinery, directly contributing to greater automation in heavy equipment sectors. Videos: this https URL
This is the author's version of the work. It is posted here by permission of the AAAS for personal use, not for redistribution. The definitive version was published in Science Robotics on 11 November 2026, DOI: this https URL - [430] arXiv:2611.18367 [pdf, html, other]
-
Title: DECODE: Decomposable Environments for COntrolled Sequential-Decision EvaluationDiego Silva Prado, Bruno Caldas de Souza, Tiago Amaral Penteado de Jesus, Ricardo Guimarães Herrmann, Cristiano Breuel, Cristiano Bevitori Maffia De OliveiraSubjects: Machine Learning (cs.LG); Risk Management (q-fin.RM); Machine Learning (stat.ML)
When a sequential decision system falls short of an optimum, its score does not reveal which modifications would improve its decisions. DECODE is a benchmark for testing such modifications through controlled experiments on planning and available information. Its consumer-credit collections generator provides five partially observed environments that share known dynamics matrices and vary a memory parameter and process noise. The experiments change selected inputs or planning rules while keeping the rest of the solver fixed. Each comparison specifies the solver's state inputs, response inputs and planning horizon. On the deterministic base environment, which was selected for a large gap between six-month and one-month planning with the hidden state, extending planning from one to six months increases the scores of two gradient-boosted solvers by 15 to 17 percentage points of the exact optimum, computed with the hidden state and borrower traits. Even at this horizon, the best graded solver, whose repayment response is fitted on logs, remains 15.6 points below that optimum. In a specified intervention sequence, three changes to its state inputs and their evaluation during search lower its score by 6.3 points, and replacing its fitted repayment model with the true response, evaluated with hidden borrower traits, then raises it by 15.6, a net recovery of nine of those points. Because the intermediate effects are conditional on the repayment model and the order of the changes, they do not uniquely assign the gap to separate causes. Within these limits, DECODE makes hypotheses about solver performance testable through specified interventions under known dynamics and explicit information conditions.
- [431] arXiv:2611.18369 [pdf, html, other]
-
Title: How rigid is maximum cardinality search? Stability and locality of MCS orderings under edge updates, verified in LeanComments: 14 pages, 1 table. Code and proofs: this https URL . Palomar registry: PALOMAR-2026-10-04-000009Subjects: Data Structures and Algorithms (cs.DS); Logic in Computer Science (cs.LO)
For which sets F of edge modifications do a graph G and the modified graph share a maximum cardinality search (MCS) ordering? We prove that they do whenever, along every proper common MCS prefix, each unchosen vertex's neighbour count in the second graph exceeds its count in the first by zero or one. In particular, inserting or deleting any matching, or flipping a single edge, preserves a common ordering, and under the same hypothesis every common MCS prefix extends to a full common ordering without backtracking. Neither part of the hypothesis can simply be dropped: inserting a triangle, a P_4 or a 2P_3 shows that the +1 bound cannot be relaxed to +2, and a disjoint mixed insert/delete pair shows that the one-directional condition is needed. For the canonical ordering (smallest-index tie-breaking) under a single-edge update, an insertion can first change the ordering only strictly between the endpoints' positions, and a deletion only at the later endpoint; when no break fires, the ordering is unchanged. A boolean-adjacency executable refines the abstract probe-then-repair update. All of these results, including the four obstructions, are proved in Lean 4 over Mathlib, re-checked by independent kernel checkers, and registered in the Palomar registry. On paper we further show that the first break of an insertion is located in O((deg(v)+1) log n) time, independent of the window length. We claim no worst-case asymptotic speedup over recomputation.
- [432] arXiv:2611.18370 [pdf, html, other]
-
Title: f-Divergence Self-Play for Tabular Anomaly Detection via Large Language ModelsComments: Accepted at ICML 2026. 28 pages, 7 figuresSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Anomaly detection in tabular data poses significant challenges due to heterogeneous feature types-mixing numerical, categorical, and textual attributes, which complicate learning meaningful representations of normality. Recent work has applied large language models (LLMs) to this problem by serializing table rows as text sequences, yet these approaches rely on one-shot supervised fine-tuning that offers limited signal to tighten the model's description of normality. We propose DiSPaT, a self-play fine-tuning framework that strengthens the model's understanding of normal data. Building on the theoretical foundation of f-divergence minimization, we derive a tight approximation connecting our training objective to reducing the distributional gap between real normal data and model-generated samples. DiSPaT operates through an alternating optimization: at each iteration, the current policy generates synthetic samples that serve as pseudo-anomalies, while a critic discriminator learns to distinguish these from real normal samples; this signal drives policy updates that progressively align the model distribution with the true normal-data distribution. Extensive experiments on diverse benchmarks demonstrate that DiSPaT consistently outperforms prior LLM-based methods, deep learning approaches, and classical unsupervised detectors for tabular anomaly detection.
- [433] arXiv:2611.18371 [pdf, other]
-
Title: Evaluating Human Sensitivity to Gender Biased AI-Generated Content through Human Assessment of Google's Autocomplete FunctionTamara Serrano, Nikoleta Pantelidou, Raquel Montero, Natalia Moskvina, Paolo Morosi, Evelina LeivadaComments: 11 pages, 2 figures, 1 tableSubjects: Computation and Language (cs.CL)
Large Language Models (LLMs) are increasingly present, but their effects on human users remain underexplored. We tested whether AI-attributed information influences views regarding gender-related statements, a domain where biases are pervasive. Spanish-speaking participants recruited through Prolific evaluated biased statements derived from Google's autocomplete function and human-authored, non-biased gender-related facts. Participants were either informed that some of the material was AI-generated or received no information about its origin. AI attribution did not affect reaction times or agreement judgments. This result suggests that AI-generated content does not automatically raise credibility concerns and may be evaluated with less scrutiny compared to human-produced content.
- [434] arXiv:2611.18372 [pdf, html, other]
-
Title: Vision-Language-Action Models: A Life-Cycle Survey of Design, Training and DeploymentSuvarna Kadam (Witsense AI), Soran Parsa (University of Huddersfield)Comments: 43 pages, 9 figures, 11 tablesSubjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET); Machine Learning (cs.LG)
Vision-language-action (VLA) models adapt web-pretrained vision-language models to output robot actions, so that web knowledge carries over to physical control. In under 4 years the field has moved from RT-2, the first VLA, to open models that fine-tune on a single consumer GPU and to proprietary generalists that clean unseen homes. This survey reviews 88 works from December 2022 to October 2026, mainly on manipulation and navigation, plus humanoid control, world-action models and evaluation. We follow a VLA through its life cycle, from model design to training and deployment, along seven axes. Model design covers (a) inputs and memory, what the model observes; (b) architecture, how its pipeline is built; (c) action output, how it produces motor commands; and (d) intermediate predictions, what else it predicts, from reasoning steps to future observations. Training covers (e) pretraining data, the mix of robot, web, human and simulated data, and (f) post-training, adaptation to the target robot. Deployment covers (g) real-time execution, keeping pace with the robot. We analyse 22 models on these axes and report their training data, model size, inference speed and hardware. Three trends emerge: (i) training data has broadened, from about 10,000 hours of teleoperated robot data ($\pi_0$) to 23,000 community-collected episodes (SmolVLA) and 20,000 hours of egocentric human video (GR00T N1.7); (ii) model sizes have split: most recent models have 3-7 billion parameters, while a new group under one billion runs on consumer GPUs; and (iii) generalist VLAs are emerging that handle both manipulation and navigation, such as UniVLA and Qwen-VLA. We also compare VLAs with world-action models, which predict future observations as well as actions. We review evidence that some VLAs with near-perfect benchmark scores fail when objects or the camera move slightly, and open up further research questions.
- [435] arXiv:2611.18375 [pdf, html, other]
-
Title: AI-Based Multi-View Echocardiographic Image Synthesis using Dynamic Shape ControlComments: 5 pages, 2 figures, accepted for publication in the Proceedings of the 2026 IEEE International Ultrasonics Symposium (IUS 2026)Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Although automated methods have improved echocardiographic workflows, progress remains constrained by the need for large, well-annotated datasets. As the development of clinical databases is often limited by ethical and logistical challenges, synthetic data generation has emerged as a promising alternative. However, existing approaches commonly rely on left ventricle shape-based supervision or are tailored to specific cardiac views, limiting the generation of additional standard and unconventional imaging planes and the preservation of multi-view consistency across cardiac chambers. To address these limitations, we propose a novel framework for shape-controlled synthesis of multi-view echocardiographic sequences. First, multi-view cross-sections are extracted from dynamic whole-heart meshes. Ultrasound-like appearance is then introduced through a two-stage AI-based pipeline. In Stage 1, an unpaired GAN-diffusion model translates these cross-sections into realistic echocardiographic frames. As this translation is performed independently at the frame level, Stage 2 employs masked video reconstruction to improve temporal coherence and cardiac motion. Evaluation demonstrated high frame-level realism for both stages, while the integration of Stage 2 substantially improved video-level motion quality and temporal consistency. Overall, the proposed framework enables the generation of realistic multi-view echocardiographic sequences from diverse cardiac models, providing a scalable source of annotated data for dataset augmentation and the development of multi-view cardiac assessment methods.
- [436] arXiv:2611.18382 [pdf, html, other]
-
Title: Where Tree Ensembles Decide: Structural Local Explanation with Decision Predicate GraphsComments: Code, data and results: this https URL. Supplementary material attached as an ancillary fileSubjects: Machine Learning (cs.LG)
Tree ensembles are among the most accurate models for tabular data, yet the computation behind a single prediction is spread over hundreds of decision rules. Feature attributions and counterfactuals explain such predictions through the influence of features or a change of the input, leaving open which of the thresholds applied by the trees a prediction responds to, while structural explanations compress these rules and lose the trees behind them. To address this gap, we propose a local Decision Predicate Graph (local DPG), which gathers the rules executed for a sample into a verifiable record of the computation, with a class support that reproduces every explained prediction. From these rules we select the critical predicate, the executed predicate whose crossing is estimated to move the model most strongly toward a chosen class, which offers the user a concrete threshold and the trees that share it, without an external explainer. Across 16 datasets and four ensemble families, it separates differently predicted neighbours better than random and nearest-threshold predicates. A single crossing recovers the true class in 46-69% of wrong predictions, against 29-50% when the same rule is applied to the feature ranked first by TreeSHAP, a measure close to the selection criterion, and on planted faults it finds the faulty threshold in 46-71% of the resulting errors, against 6-21% at random. Although training-free edits of these predicates did not reduce the error on unseen data, the critical predicate complements attributions and counterfactuals by naming where a prediction can be changed.
- [437] arXiv:2611.18385 [pdf, html, other]
-
Title: Voice-Based Conversational Agents as AI Care Assistants: Evaluating Personalization, Intervention Delivery Fidelity, and Privacy Expectations in Older Adults' CareThalia Viranda, Shupeng Wang, Jia Bo, Yuchen Yang, Hongjiao Zhang, Andrea Cuadra, Daniel A. Adler, Walter Boot, Sara Czaja, Deborah Estrin, Tanzeem ChoudhurySubjects: Human-Computer Interaction (cs.HC)
Generative AI enables voice-based conversational agents to play more active roles in older adults' care, including personalizing support and operating alongside clinicians and care partners. This shift creates two challenges: how to personalize support while faithfully delivering evidence-based interventions, and how to manage personal-data use and sharing across AI and care stakeholders. We designed and evaluated ARC, a voice-based conversational agent that operationalizes multidomain lifestyle intervention through sessions with predefined goals, participant context, and evidence-based components. In a 3-day remote study, 80 older adults completed 473 sessions across six session types (9,183 turns; averaging 19.4 turns and 6 minutes per session). Participants reported favorable usability, acceptability, and use intentions. System-level analyses showed strong domain-level but modest participant-specific contextual reuse, no significant personalization-fidelity association, and fidelity limitations centered on incomplete delivery of intervention components rather than off-protocol drift. Within care relationships, expectations for personal-data use and sharing varied by data type and recipient. Our findings support feasibility of conversational delivery of multidomain lifestyle intervention, offer methods and system-design insights for longitudinal evaluation, and motivate an evaluation agenda for generative AI health interventions: user-experience and outcome measures should be complemented by evidence of whether intended intervention processes were delivered as designed.
- [438] arXiv:2611.18390 [pdf, html, other]
-
Title: CS-MPC: Contraction-Certified Sampling-Based MPC for Nonlinear Systems Under UncertaintySubjects: Systems and Control (eess.SY); Robotics (cs.RO)
We present contraction-certified sampling-based model predictive control (CS-MPC) for tracking and safety of nonlinear robotic systems under parametric uncertainty, disturbances, and safety constraints that are revealed online. CS-MPC uses a contraction-certified tracking feedback inside every sampled rollout, which concentrates the rollouts around the reference, and an execution layer that verifies the applied input using a quantile-tightened discrete-time control barrier function condition. We derive a verification procedure for nonlinear contraction of the feedback-closed dynamics, and we prove a finite-sample stochastic stability bound and a mission-level probabilistic safety guarantee. We validate CS-MPC in simulation on a rover, a quadrotor, and an asteroid landing. When compared with standard model predictive path integral (MPPI) control, CS-MPC attains the same tracking error with 7 to 20 times fewer rollouts for the rover, and a 2 to 4 times lower tracking error for the quadrotor, with no safety violations in 200 Monte Carlo missions.
- [439] arXiv:2611.18391 [pdf, other]
-
Title: Color Refinement Distinguishes Connected Rooted Outerplanar GraphsSubjects: Discrete Mathematics (cs.DM); Data Structures and Algorithms (cs.DS); Combinatorics (math.CO)
Color refinement, also known as the $1$-dimensional Weisfeiler-Leman algorithm ($1$-WL), is a classical near-linear-time heuristic for graph isomorphism, which also captures the expressive power of message-passing graph neural networks. A long line of work asks on which graph classes $1$-WL, or its higher-dimensional variant $k$-WL, identifies every graph; the least such $k$ is the WL-dimension of the class. Planar graphs have WL-dimension at most $3$ (Kiefer, Ponomarenko and Schweitzer, 2019), and whether $2$-WL already distinguishes all planar graphs is a central open problem. We bring new insight on this question for outerplanar graphs, whose WL-dimension is $2$. We show that color refinement, i.e. 1-WL, distinguishes connected rooted outerplanar graphs: two non-isomorphic such graphs receive different stable colorings once a single vertex, the root, is individualized. The root is necessary, as there are non-isomorphic $2$-connected outerplanar graphs that color refinement does not distinguish. Our result extends to pseudo-rooted graphs, in which the stable coloring merely contains a singleton color class, and to non-connected graphs whose components all contain such a class. Consequently, the $(1,1)$-WL test of Rattan and Seppelt (2023) distinguishes all outerplanar graphs. We proceed by inversion: we reconstruct a connected rooted outerplanar graph, up to isomorphism, from its color graph, the weighted quotient of the graph by its stable color classes, together with its root color. The reconstruction runs in time linear in the output size and is canonical, which yields a near-linear-time canonization algorithm for these graphs. Our main tool is a dichotomy lemma, formalized in Rocq: in a rooted outerplanar graph with an equitable coloring, every block either receives pairwise distinct colors or admits a color-preserving mirror symmetry fixing its vertex closest to the root.
- [440] arXiv:2611.18392 [pdf, html, other]
-
Title: VelServe: Taming Decode Jitters in LLM Inference with Tail-Calibrated SchedulingSubjects: Artificial Intelligence (cs.AI)
Interactive LLM services must meet latency service-level objectives (SLOs) for the time to first token (TTFT) and the time between tokens (TBT), as streaming and soft real-time applications require predictable token delivery. We identify decode jitters in both aggregated and disaggregated serving. A jitter delays every request in the batch, and long responses span hundreds of steps, so even rare overruns cause many SLO violations. In aggregated serving, SLO-aware schedulers set each step's prefill budget until its predicted latency just meets the TBT target, leaving no room for mispredictions or host stalls. Disaggregation meets average TBT targets at a cost in capacity, yet the steps that admit new requests still run late.
We present \sysname, a scheduler that calibrates how much prefill work each decode step carries using the measured tail of its own iteration latency. The SLO determines how rare a late step must be, and \sysname leaves enough headroom in each iteration to prevent overruns and meet the target decode latency. This headroom slows prefill, so \sysname serves the shortest waiting prompt first. When TTFT deadlines grow with prompt length, the long prompts it delays have the most time to spare. Under disaggregation, it spaces admissions so that no request waits on more slow admission steps than its P99 target allows. On three attention families and varying workloads, \sysname sustains 2.4--2.9$\times$ the SLO goodput of the strongest baseline on Qwen3-8B and DeepSeek-V4-Flash, and at least 4.4$\times$ on DeepSeek-V2-Lite. With prefill and decode disaggregated on eight GPUs, \sysname meets the SLO for more than twice as many requests as state-of-the-art baselines. - [441] arXiv:2611.18393 [pdf, html, other]
-
Title: WeatherDreamer: Learning a World Model from Multimodal Earth ObservationsSubjects: Machine Learning (cs.LG); Atmospheric and Oceanic Physics (physics.ao-ph)
Machine learning weather models now rival operational global forecasts but still start from analyses produced by separate, expensive data assimilation (DA) systems, which for global models are typically updated every six hours. Observation-to-observation models avoid this dependence, but most omit the recursive filtering cycle that lets DA systems carry a background state through sparse, irregular observations. We introduce WeatherDreamer, a world model that learns this cycle end-to-end from 18 global observing systems, without reanalysis. It infers a latent state from observations, carries it forward hour by hour with learned dynamics, corrects it each hour with new observations, and can predict what any of its observing systems would measure at any location. At held-out stations, its ensembles beat NOAA's Global Ensemble Forecast System (GEFS) for temperature, dew point and wind out to 12 h and are better calibrated than GEFS and the ECMWF ensemble. Precipitation, cloud-top height and solar irradiance, never inputs at inference, are inferred with skill matching or exceeding operational products at short leads. Against the first guess of NOAA's Gridpoint Statistical Interpolation (GSI) DA system, WeatherDreamer's short-range departures are within 10% of GSI's at surface stations and within 25% for moisture-sensitive microwave channels, but up to five times larger for temperature-sounding channels. Never given observation errors, its learned update assumes errors close to known instrument and retrieval errors of microwave sounders and scatterometer winds. WeatherDreamer also separates viewing geometry from atmospheric state, recovering 83-96% of observed limb darkening and producing seamless nadir-view imagery. Producing an hourly global analysis in about 1.5 minutes on one GPU, WeatherDreamer shows that assimilation and forecasting can be learned together directly from observations.
- [442] arXiv:2611.18395 [pdf, html, other]
-
Title: Smart Home Agents Make Capable but Unpredictable Grid OperatorsComments: Pre-print. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessibleSubjects: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
AI-enabled smart homes are becoming more common across the grid edge. However, there has not been much work on failure response. This area is critical because smart homes operate on the grid edge where failures can cascade down. If a power outage happens within a smart home, a user has to manually troubleshoot their smart home to include tasks such as power load balance and device recovery. We conduct the first resilience analysis of smart home agents across a full power outage, providing a suite of outage simulations in a live environment that measures resilience before, during, and post power outage. We show that smart home agents can act as grid operators in power outages and identify several new failure categories highlighting the current limitations across all of our models. These findings establish the groundwork for smart home agents as grid operators to increase overall resilience on the grid edge.
- [443] arXiv:2611.18396 [pdf, html, other]
-
Title: Deep Neural Networks Training Efficiency: Does Grokking Modify its Dynamics?Subjects: Machine Learning (cs.LG)
We study whether grokking, the delayed emergence of out-of-distribution generalization after apparent overfitting, alters training efficiency in deep neural networks, where efficiency is measured as accuracy per unit energy. Using two canonical settings, modular arithmetic and MNIST, we analyze two neural network models under 16-, 32-, and 64-bit precision and use Orthograd to test whether grokking produces noticeable deviations from the typically decreasing dynamics of training efficiency. We consider both short training runs of up to 500 epochs and long training runs of up to 20,000 epochs, and in both cases identify grokking as a sharp improvement in test accuracy after a sustained train-test gap. In the short-horizon experiments, applying Orthograd at progressively stronger levels during early training does not induce grokking, and efficiency remains smoothly decreasing. In the long-horizon experiments, Orthograd induces grokking, but the efficiency curves remain approximately monotonic and show no appreciable local increase at the grokking transition. These results suggest that, in our setting, grokking does not materially alter overall training-efficiency dynamics.
- [444] arXiv:2611.18397 [pdf, html, other]
-
Title: Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game AgentRun Luo, Kuan Zhang, Qi Deng, Yukun Chen, Zhihao Yang, Yufei Zhang, Renrong Hu, Yi Zhang, Guiyu Ma, Tianyu Fan, Longze Chen, Hao Sun, Jiaming Li, Jiaming Yang, Xiaohao Liu, Tao Tan, Minzheng Wang, Hangtian Jia, Shuai Yuan, Lvdi Wang, Xiao Liu, Chao Ma, Yuting Zou, Yinhe Zheng, Xiaobo Xia, Dawei Jiang, Shaoqing RenSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
We introduce Gamer, a generalist agent framework built for long-horizon tasks in complex interactive environments. Its end-to-end architecture combines perception, memory management, and multi-step reasoning. A shared training and deployment pipeline supports interaction with both conventional 2D graphical interfaces and complex 3D simulated worlds. We first continue pre-training on a dataset constructed from tens of thousands of hours of device-level keyboard and mouse interaction. These trajectories span simulated games and embodied environments, allowing the model to learn environmental dynamics across both physical and virtual domains. Sparse-thinking instruction tuning then substantially reduces redundant computation while preserving the reasoning needed for complex decisions. We train the policy for long-horizon interaction with a stable multi-turn reinforcement learning method that addresses finite context windows, accumulated rollout errors, and sparse rewards. For real-time interaction, a cloud-edge inference engine serves numerous simulated game environments concurrently, supports unbounded-context execution, and keeps online inference latency low. Across multiple challenging 3D open-world games, Gamer demonstrates solid generalization and operational efficiency comparable to skilled players. It also achieves competitive results on a range of conventional 2D game benchmarks. These results suggest that even repetitive keyboard and mouse trajectories can support substantial gains from scaling data and learning algorithms. Training on these trajectories helps Gamer learn complex skills more steadily in environments with both physical and digital properties. This approach offers a path toward generalist agents that can interact across digital and physical worlds.
- [445] arXiv:2611.18399 [pdf, html, other]
-
Title: A Framework for Cost-Preserving Rounding with Applications to Laminar Thin TreesSubjects: Data Structures and Algorithms (cs.DS)
In the context of finding thin trees, like laminarly-thin trees or globally-thin trees, it turned out to be challenging to round a fractional solution to an integral one without losing cost, while only having a constant-factor violation of the thinness constraints. We provide a general framework for this problem, in a much broader context, that reduces the problem to finding a certain type of rounding algorithm with only an approximate prize guarantee. We show that, up to a change in constants, the rounding problem with approximate prize guarantee that we use is equivalent to the original question. As an application, we obtain a polynomial-time algorithm that finds a laminar constantly-thin tree of cost no higher than the one of the natural linear programming solution, thus obtaining the first result without cost loss of this type (even for the special case of chain-constrained trees). Moreover, our framework gives a new perspective on how to approach the problem of finding globally-thin trees without cost loss.
- [446] arXiv:2611.18409 [pdf, html, other]
-
Title: Fuse-PINN: Combining Finite Element Methods and Physics-Informed Neural Networks to enable massively parallel simulations of non-linear PDEsStanislas Pamela, Nicolas Pailliez, Emmanuel Franck, Victor Michel-Dansac, Laurent Navoret, Vignesh GopakumarComments: 9 pages plus Appendix, 11 figures, 1 table. Rejected at ICLRSubjects: Machine Learning (cs.LG); Plasma Physics (physics.plasm-ph)
Research in the domain of Physics-Informed Neural Networks applied to PDE modelling has advanced significantly in recent years, providing a wide ecosystem of novel methods. However, one frequent limitation is the inability of PINNs to scale out on High-Performance Computing architectures to address more ambitious challenges, particularly non-linear physics models with turbulent dynamics in large spatial domains. Typical studies restrain demonstrations to toy models with simple physics equations, few variables, and rarely any complex non-linear dynamics evolved over long timescales. Part of the AI community has even adopted the idea that "PINNs don't scale". This work proves otherwise. We present a mathematically robust formulation of Finite-Element-Method for Neural Networks (FEM-NN) implemented into a comprehensive, modular framework: Fuse-PINN. This formulation provides the basis for a sclable parallelisation of Physics-Informed Neural Networks on GPU-based HPC systems. Furthermore, an Energy Natural Gradient Descent optimiser is implemented to allow high-precision convergence. Fuse-PINN is demonstrated on physics models relevant to Magnetic Confinement Fusion, including non-linear convective models with fine-scale dynamics. Near-perfect scaling is achieved on multi-GPU runs with increasing problem size.
- [447] arXiv:2611.18412 [pdf, html, other]
-
Title: LitArc-Bench: Disentangling Text-Specific Reader-Response Dynamics from Position and Lexical AffectComments: 13 pages, 2 figures, 12 tablesSubjects: Computation and Language (cs.CL)
When model predictions track an unfolding human affect trajectory, alignment may reflect narrative position, affective words, or text-specific reader-response dynamics. We introduce LitArc-Bench to separate these explanations: 39 Chinese-language literary excerpts segmented into seven checkpoints, with 430 questionnaires from 81 adults in two classroom groups (9,020 VAD ratings). We compare three future-blind model runs with leave-one-text-out position priors, lexical affect, and human-human references. Position alone reaches r = .470 for arousal, making raw alignment insufficient. Yet all six V/A transition advantages over a joint position-lexicon control have text-bootstrap intervals above zero. Cross-fitted models match the correct human text better than 38 mismatched texts in all nine model-dimension cells (p < .0001); a text-invariant negative control returns chance (p = 1). On human transitions at least 0.5 points against the position trend, models recover 69-76% of valence directions and 53-68% for arousal and dominance. Under crossed participant-text resampling, all nine identity intervals remain above zero; joint-control advantages are robust for valence and directional but less certain for arousal. High shape agreement still coexists with scale bias. LitArc-Bench identifies which components of literary reader response are recoverable beyond shallow explanations and why shape, calibration, and human variability require separate evaluation.
- [448] arXiv:2611.18417 [pdf, html, other]
-
Title: CoPre-MoE: Consolidating and Preserving MoE for Multimodal Continual Instruction TuningSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Multimodal continual instruction tuning requires multimodal large language models to sequentially acquire new capabilities without catastrophic forgetting. Mixture-of-experts (MoE) architectures offer a promising way to mitigate task interference through expert specialization and combination. However, while existing MoE-based methods avoid overwriting shared parameters by allocating separate experts for each task, this progressive expansion incurs severe cross-task redundancy and linear parameter growth. Specifically, experts across tasks often learn overlapping functions alongside task-specific capabilities. For instance, ScienceQA and TextVQA both locate relevant image regions. To curb parameter redundancy without inducing forgetting, we propose Consolidating and Preserving MoE (CoPre-MoE). For each incoming task, CoPre-MoE first learns task-specific MoE to capture specialized capabilities. To eliminate redundancy and reduce parameter growth across tasks, we design a shared-and-private expert architecture. Specifically, we first evaluate functional overlap to selectively consolidate overlapping experts across tasks into a fixed-capacity generic expert pool. We then factorize the remaining task-specific experts into shared projection bases and compact task-specific cores within a private expert pool. To mitigate forgetting, we construct the router by pairing each expert with learnable retrieval keys, where historical keys remain frozen during continual updates while newly added keys are regularized with an orthogonality penalty. At inference, input-dependent routing dynamically combines both pools without test-time task labels. Extensive experiments demonstrate that CoPre-MoE achieves superior performance while substantially reducing parameter overhead.
- [449] arXiv:2611.18419 [pdf, html, other]
-
Title: LECAT: Measuring LLM Category Learning from Pretraining DataComments: 21 pages, 4 figures, 8 tablesSubjects: Artificial Intelligence (cs.AI)
Humans know that airplanes are fast, that this extends to red airplanes, but that it does not hold for broken airplanes. How do LLMs acquire category knowledge that humans learn and apply effortlessly? We introduce LeCat, a two-part framework to analyze LLM's Learning of Category knowledge. First, we use LeCat to evaluate models on their ability to learn category knowledge, distinguishing between three cases: memorization of in-domain category knowledge, generalization to unseen instances, and ability to flexibly use knowledge in appropriate contexts. Then, we use LeCat to trace how this knowledge can be learned in pretraining from text that may encode category knowledge in various ways, such as verbatim repetition or exceptions, which we refer to as forms of "knowledge expression". We first assess open-source pretrained models using 199 curated category-feature pairs and analyze their training data. Our findings suggest that strengthening context and category-feature co-occurrence likelihood (PMI) improve learning for larger models and verbatim rule-like statements diminish flexible category learning. To systematically isolate these effects, we then train a 68M model from scratch with precise control over the categorical information each sees in their training data, confirming that verbatim repetition reduces flexible learning but mixture of diverse knowledge expressions lead to more robust category learning. Together, our findings highlight that while contextual flexibility is persistently difficult for models, how category knowledge is expressed in pretraining greatly impacts models' acquisition of robust category knowledge.
- [450] arXiv:2611.18422 [pdf, html, other]
-
Title: The State Is Not the Bottleneck: A Paged-State Gated-DeltaNet Tile on a Tri-Mode Multicast Mesh, Bit-Exact from RTL to SiliconComments: 12 pages, 5 figures, 6 tables. The technical supplement is included as an ancillary file. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessibleSubjects: Robotics (cs.RO); Hardware Architecture (cs.AR); Machine Learning (cs.LG)
Linear-attention language models replace a growing key-value cache with a recurrent state of fixed size. On an edge FPGA, this state can exceed the on-chip memory and must be paged to external memory. This paper asks whether such paging dominates the cost of decoding (generating tokens). We built a hardware tile that computes one complete layer of a custom 369-million-parameter hybrid Gated DeltaNet model on an Agilex 5 FPGA and pages its recurrent state. The model state needs 6.2 MiB, whereas the device has 0.87 MiB of block memory. Paging adds 26.0 MB of memory traffic per decoded token, at most 11.2 percent of the calculated total. On the board at 75 MHz, with synthetic weights, one token takes 11.15 ms in this layer. The matrix-vector products with the weights take about 62 percent of this time and the state phase, including the paging, 13 percent. The tile is one node of a multicast mesh of tri-mode processing elements. We measure the error of the fixed-point arithmetic against floating point and check that the hardware agrees bit for bit with a fixed-point reference. In simulation, the state logic agrees with the reference on all 18 recurrent layers. On the board, six builds pass all 30 runs of 64 tokens and two runs of 1,024 tokens. For this model and memory interface, the delivery of the weights limits the decoding performance, and the recurrent state does not. State paging is therefore practical for such a model on an edge FPGA.
- [451] arXiv:2611.18428 [pdf, html, other]
-
Title: Grid-Forming AI Data Centers: Coordinated Grid Support Control from IT Loads and Energy StorageSubjects: Systems and Control (eess.SY)
To exploit the grid-support capability of AI data centers (AIDCs), this paper proposes a grid-forming (GFM) control scheme which coordinates the IT loads and energy storage (ES) in the AIDC. First, the framework of a grid-connected AIDC is presented, with the multi-stage power conversion from the grid to the IT loads presented. A dual GFM scheme is then proposed for the back-to-back converters to provide grid support through coordinated power adjustments from IT loads and ES. In addition, an aggregated power response model is established to characterize the IT-load flexibility considering multiple IT servers executing diverse tasks. Time-domain simulations of a 100 MW grid-connected AIDC demonstrate that the proposed control scheme can effectively support the grid.
- [452] arXiv:2611.18429 [pdf, html, other]
-
Title: Do Underwater Image Quality Metrics Agree with Human Perception? A Benchmark with Human Choices, Simple Edits and Generative DriftComments: Accepted at the 6th WVAQ Workshop, WACV 2027. 18 pages including appendixSubjects: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
Underwater image enhancement is challenging to evaluate due to the lack of clean reference images, so researchers rely on no-reference image quality assessment (NR-IQA) metrics. However, it remains unclear whether these metrics agree with human judgments. We evaluate 73 handcrafted underwater, learned underwater, vision-language and general-purpose metrics against human preferences on five references: four published opinion-score sets and a new paired-comparison study. For each metric, we measure how often it selects the same image as people when comparing versions of the same scene. We further test whether simple image edits and small adversarial perturbations raise metric scores above the enhancement people preferred, and whether scores decrease as repeated generative edits add noise. Our results show that no metric consistently matches human preferences. The best worst-case Cohen's kappa over the five references is only 0.23 (Q-Align), while two raters in our study agree with a median kappa of 0.41. The widely used UCIQE often favors edits that people reject. On a Nano Banana Pro replica chain that should leave the image unchanged, only 22 of 73 scores decrease at every checkpoint, while noise grows by 2.5 times. We also introduce TrustUIQA, a benchmark that combines the human references, our study choices, the edited, attacked and generated images, and the scores of every metric on every image. The code is available at this https URL and TrustUIQA at this https URL.
- [453] arXiv:2611.18434 [pdf, html, other]
-
Title: Towards Explainable Sleep Stage Classification using Single-Channel ElectroencephalogramComments: 8 pages, 4 figures. Accepted at AgIA4BioHealth Workshop, IEEE BIBM 2026, Dallas, Texas, December 1-4, 2026Subjects: Artificial Intelligence (cs.AI); Signal Processing (eess.SP)
Automated sleep staging has achieved strong bench- mark accuracy; however, clinical adoption remains limited due to a lack of interpretability. In this work, we fill the research gap by extending the SleepXAI sleep staging framework to systematically compare five different XAI methods. Further, attribution fidelity is evaluated using Spearman rank correlation with AASM-defined EEG frequency bands, serving as clinically grounded reference signals. In this paper, we focus on single- channel Electroencephalogram (EEG), which is more amenable to practical clinical deployment compared to full polysomnography setups. Experiments are conducted on the Sleep-EDF benchmark dataset and the DREAMT dataset. In DREAMT, we stratified subjects based on OSA into four Apnea-Hypopnea Index (AHI) severity groups. We observed that the SleepXAI model achieves strong classification performance on both datasets. Among the evaluated methods, Integrated Gradients (IG) consistently achieves the highest mean attribution fidelity across all five sleep stages on Sleep-EDF and across all AHI groups on DREAMT, outperforming all other evaluated methods under the Spearman rank correlation metric. These findings highlight the importance of attribution method selection as a key factor for ensuring reliable and clinically interpretable automated sleep staging systems.
- [454] arXiv:2611.18436 [pdf, other]
-
Title: Efficiency Gains in Social Science Research: A Prospective Silicon Sampling Protocol for Pre-Treatment and Code OptimizationComments: 14 pagesSubjects: Computers and Society (cs.CY)
This study evaluates the methodological potential of using large language models (LLMs) to generate synthetic datasets prior to real-world data collection, thereby streamlining subsequent processing and statistical analysis. The approach is applied to a survey conducted within the France Life Imaging (FLI) network, leveraging three distinct LLMs (Sonnet 4.6, GPT 5.3, and Kimi 2.6) to simulate responses and prepare data processing workflows. The findings demonstrate significant time savings in code development (85% to 90% reuse rate) and highlight the conditional validity of this preprocessing method, which depends on the thematic predictability of the models. While synthetic data effectively captures collective perceptions, it struggles with specialized or contextualized practices, reflecting limitations documented in the silicon sampling literature, such as homogenization and reduced variance. This work extends the Phase 0 concept (Trojacka et al. 2025) from instrument validation to a full-fledged methodological preprocessing step, paving the way for scalable applications in survey research.
- [455] arXiv:2611.18437 [pdf, other]
-
Title: Accounting for Trust and Safety: Bounding Content-Moderation Quality from Records You Already HoldComments: 23 pagesSubjects: Computers and Society (cs.CY)
Platforms use expert audits to measure content-moderation quality. This paper develops a formal method for checking those claims using records platforms already maintain, such as appeal reversals and confirmed reports. These records provide one-sided bounds: they can show that a stated quality target was missed, but cannot prove that it was met. The paper presents the theory, limitations, constructed examples, and a prospective protocol for future evaluation. It reports no analysis of Meta operational data, no experiment, and no production result.
- [456] arXiv:2611.18438 [pdf, html, other]
-
Title: Toward On-Device Markerless 6DoF Tracking of a Physical Manipulative on a Mixed-Reality Headset: A Preliminary StudySubjects: Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
AR-Classroom uses physical models and virtual overlays to help learners explore spatial transformations. To align these overlays on a standalone headset, we developed a markerless six-degree-of-freedom(6DoF) tracker for a LEGO airplane on the Meta Quest 3. A compact YOLO model trained on synthetic Unity renders predicts 20 keypoints from dynamic RGB crops resized to 128 by 128 pixels. A perspective-n-point solver estimates the airplane's pose from these predictions and the corresponding 3D landmarks. In a paired evaluation on 1,000 synthetic validation images, cropping guided by ground-truth boxes increased the detection rate on annotated images from 32.5 to 46.2 percent. Mean keypoint F1 increased from 42.9 to 44.6 percent, with lower keypoint precision in the cropped view. In preliminary headset tests, 2 Hz was the highest tested tracking target at which the application maintained approximately 72 frames per second. A headset demonstration illustrates virtual overlay alignment with the physical airplane before and after a manual rotation. Quantitative 6DoF pose accuracy on real headset footage will be measured in future work.
- [457] arXiv:2611.18439 [pdf, html, other]
-
Title: Between Obligations and Evidence: A Participatory Study of Translating the EU AI Act into Technical RequirementsHelma Torkamaan, Masoud Mansoury, Yanan Xin, Oscar Oviedo Trespalacios, Roberta Savella, Andrej Zwitter, Maria Heuss, Huy Son Nguyen, Marie-Therese Sekwenz, Moyi Li, Roberta Tamponi, Megha KhoslaSubjects: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Software Engineering (cs.SE)
The EU AI Act imposes technical and organizational requirements for high-risk AI systems, but applying these requirements involves contextual choices about what to measure and what evidence supports a compliance claim. We investigate these challenges through a participatory workshop involving 32 researchers, policymakers, and practitioners from different disciplines, using realistic scenarios in healthcare, recruitment, and public transport to examine how regulatory obligations are interpreted, translated into technical requirements, and evaluated in practice. Through a thematic analysis of workshop data, we identify recurring tensions between how regulatory obligations are translated into measurable system properties, how responsibility and human oversight are exercised in practice, and how the validity of compliance evidence is affected by changing operational conditions and system uses. For example, participants highlighted how fairness metrics could conceal discrimination in recruitment, how time pressure and limited information could undermine human oversight in public transport, and how changes in patient populations could invalidate previously established evidence of system performance. Our findings suggest that technical, organizational, and evidentiary challenges are interdependent, and that establishing and maintaining compliance requires attention not only to technical assessments but also to how AI systems are used, governed, and monitored over time. We discuss implications for supporting context-sensitive and ongoing compliance.
- [458] arXiv:2611.18442 [pdf, html, other]
-
Title: CANTraj: Recovering Vehicle Trajectories from CAN Signals in Naturalistic DrivingSubjects: Cryptography and Security (cs.CR); Databases (cs.DB)
Open driving releases and usage-based insurance programs often claim to prevent privacy breaches by omitting GPS records when sharing or collecting vehicle data. However, removing GPS coordinates from driving logs does not erase the location information carried by vehicle signals. While prior route-inference attacks have demonstrated that speed, steering, and driving events can be matched to road networks to reconstruct trajectories, existing studies rely on inconsistent prior assumptions and evaluation metrics and are tested only on small, restricted datasets, making their results difficult to compare and their real-world effectiveness uncertain. We release CANTraj, a naturalistic driving dataset pairing decoded CAN events with time-aligned GNSS trajectories (3,940 drives, 454 drivers), and CANTraj-Bench for evaluating trajectory recovery (3,675 drives, 431 drivers, 1,380 hours, 85,919 km). On 776 evaluation drives from 94 unseen drivers, we compare six re-implemented attacks and benchmark baselines under explicit information-access protocols, paired comparisons, and metrics incorporating failures. Experiments show that combining steering signals with road-network constraints yields the largest gain: With a known start and heading, adding steering to speed raises the online map filter's five-minute route recovery from 56% to 87%, with a median average displacement error of 43 m versus 921 m for map-free dead reckoning. Offline beam search recovers 88% of test routes with the heading and 81% without it. Because methods fail on different windows, we also evaluate an LLM harness that selects among candidate routes on 659 five-minute windows; its best configuration recovers 79% of routes on test and 83% on few-shot, outperforming the best baseline on test and matching it on few-shot. Data and code at this http URL and this http URL.
- [459] arXiv:2611.18443 [pdf, html, other]
-
Title: HerAgent: A Block-Level Checkpointing Harness for Coding-Agent Environment SetupComments: 12 pagesSubjects: Software Engineering (cs.SE)
Modern coding agents rely on establishing executable software environments before they can perform software engineering tasks such as testing, debugging, and patch validation. Existing setup agents typically control progress at either single-command or whole-script granularity, making recovery either too local to remove polluted state or too coarse to preserve useful progress. We present HerAgent (Harness EnviRonment Setup Agent), a block-level harness for coding-agent environment setup. HerAgent's main contribution is semantic setup blocks: a single progress-control abstraction used for execution, validation, checkpointing, repair, and rollback. Evaluated against six prior agents on four environment-setup datasets, HerAgent improves setup success in every comparison, achieving the largest gains on Installamatic-Bench (52.0% to 85.0%) and SetupBench (46.3% to 70.4%). It also reduces setup costs by up to 94.5%, showing that block-level progress control improves reliability and efficiency.
- [460] arXiv:2611.18449 [pdf, html, other]
-
Title: Deterministic ECDSA on secp256k1: the RFC 6979 assumption, a counterexample, and the state of deployed implementationsComments: 21 pagesSubjects: Cryptography and Security (cs.CR)
RFC 6979 derives the ECDSA nonce as $k = F_x(h)$, with $F_x$ HMAC_DRBG keyed by the signing key x. We isolate the assumption its security rests on, A*: at known public key $Q = xG$, the signing transcript under nonces $F_x()$ is indistinguishable from the transcript under uniform nonces. A* suffices, and it is neither PRF security of HMAC nor auxiliary-input PRF security (AI-PRF) of $F_x$. A function $F'(x, h) = P_{Ext(xG)}(h)$ is a PRF under standard assumptions yet leaks x from one signature; the same holds for unmodified RFC 6979 over a compression function that keeps the PRF and dual-PRF properties used in proofs of HMAC, given a collision-resistance premise. A* and AI-PRF are separated in both directions. A* does not imply AI-PRF if the one-bit hidden number problem is hard. Conversely, in the standard model: any AI-PRF can be modified so that it stays one while two chosen-digest signing queries reveal x; under DDH with one public random point, or decisional square Diffie-Hellman without it, there is an AI-PRF whose key is recovered from signatures on two ordinary hashed messages; and a third function, under PRF security and DLP, violates A* while the scheme stays unforgeable. With an ideal compression function, A* follows tightly from DLP in the presence of a randomized-ECDSA signing oracle, and a nonce key independent of x removes the assumption. The secret HMAC message block is $2^{-127.65}$ from uniform on secp256k1, $2^{-32}$ on P-256 and 0.336 on brainpoolP256r1. An audit of 31 secp256k1 implementations finds that 28 sign deterministically without added entropy and that the raw-message deviation from bits2octets of libsecp256k1 before 0.2.0 survives in NBitcoin, dcrd and btcec v2. Low-R grinding needs no assumption beyond A*, and every nonzero Fourier coefficient of the low-R nonce set is below $2^{-116.9}$ unconditionally. We give test vectors for the untested $h >= n$ branch
- [461] arXiv:2611.18450 [pdf, html, other]
-
Title: PACT: A Decoder-Confidence-Guided Two-Round Receiver for Joint Channel and Phase-Noise Estimation in Sub-THz OFDMSubjects: Machine Learning (cs.LG); Signal Processing (eess.SP)
Frequency-selective fading and oscillator phase noise (PN) jointly impair coded sub-THz OFDM reception. We propose a decoder-confidence-guided two-round receiver combining a {pilot-aware complex transformer} (PACT) estimator, banded linear minimum mean-square error equalization, and neural normalized min-sum decoding. PACT jointly estimates the frequency-domain channel and effective time-domain PN through interacting complex-valued branches. Decoder-derived pseudo-pilots and continuous constellation-aware confidence guide second-round refinement through PACT inputs and attention, with parameters shared across rounds. At $E_b/N_0=30$~dB, the PACT receiver achieves approximately $70\%$ and $57\%$ lower code-block error rates than the best-performing learning-based baseline in the two benchmark settings, respectively. Frozen-weight tests further demonstrate applicability across channel/PN models, carrier frequencies, and FFT/bandwidth configurations.
- [462] arXiv:2611.18451 [pdf, html, other]
-
Title: Policy-Conditioned Koopman Representations for Parameterized Closed-Loop SystemsSubjects: Systems and Control (eess.SY)
Predicting the response of a nonlinear system across different feedback controllers requires a model that captures how the closed-loop dynamics change with the controller parameters. In a finite-dimensional Koopman model, these changes may affect both the latent dynamics and the observable coordinates used to represent them. In this paper, we propose a policy-conditioned Koopman representation framework to address this problem. This framework allows a single learned model to predict a family of closed-loop trajectories without retraining for each controller. We test the framework on four nonlinear prediction benchmarks, where the model stays below a 15% error threshold on every system, while each baseline exceeds it on at least one. As an example application, the resulting surrogate supports one-shot feedback-parameter selection: an optimization at the initial state selects a controller that then operates with fixed parameters and ongoing state feedback. On nonlinear control benchmarks including MuJoCo, the selected controllers stabilize every system with 3 to 97 times less online computation than nonlinear model predictive control. These results demonstrate how policy-dependent coordinates can support both prediction across closed-loop systems and controller selection without repeated online trajectory optimization.
- [463] arXiv:2611.18453 [pdf, html, other]
-
Title: Black-Box Uncertainty Quantification for Large Language Models via Ensemble-of-EnsemblesComments: Accepted at NeurIPS 2026. 54 pages (10 pages main text), 3 figures, 38 tables. Code: this https URLSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Reliable use of large language models (LLMs) requires uncertainty estimates that separate ambiguity intrinsic to a prompt from genuine knowledge gaps. Bayesian and deep-ensemble methods deliver this aleatoric/epistemic decomposition in principle, but are computationally prohibitive at LLM scale and require white-box access. Existing black-box methods, in contrast, summarize to a single scalar that cannot make the distinction. We introduce a two-level ensemble that bridges this gap. Stochastic decoding of a fixed prompt forms the inner ensemble; meaning-preserving semantic perturbations of the prompt form the outer ensemble. Variability is measured in a continuous embedding space, and the law of total covariance gives an exact within-/between-perturbation decomposition to estimate aleatoric and epistemic uncertainty (AU/EU); it requires only black-box sampling access. On top of this estimator, we contribute three theoretical results that ground the framework: (i) a finite-sample bias identity for nested Monte Carlo with a closed-form bias-corrected epistemic estimator; (ii) a Bayesian bridge bounding the gap between perturbation-population AU/EU and Bayesian targets; and (iii) a separation result identifying a class of failure modes that any consistency-only black-box method provably misses while perturbation-based EU detects. Across five short QA benchmarks and four instruction-tuned LLMs spanning two families and two scales, our estimator is competitive with strong white-box baselines, and at a matched generation budget it matches leading black-box methods on error ranking while decomposing AU and EU; controlled paired interventions show that AU and EU respond selectively to distinct, actionable sources of failure: prompt ambiguity and model knowledge gaps; and a long-form pilot on the NQ long-answer split shows the framework extends to claim-level factuality without retraining.
- [464] arXiv:2611.18458 [pdf, html, other]
-
Title: When to Trust Simulated Patient Responses:Support-Gated World Models for Vasopressor Decision SupportSubjects: Artificial Intelligence (cs.AI)
Clinical world models expose language-model agents to predicted patient responses, but simulations can be unreliable for poorly supported treatments or unfamiliar hospitals. We introduce Support-Gated World-Model decision support, a lightweight extension of a propose--simulate--refine agent. A recurrent dynamics ensemble predicts short-horizon physiological responses to documented norepinephrine dose categories, while a behavior model estimates observational action support. The interface withholds predictions with insufficient support or excessive ensemble disagreement and abstains when no candidate qualifies. Evaluation on eICU-CRD includes 70 hospitals and 30,700 decision transitions, with 14 hospitals held out for testing. At 62.0% factual-transition coverage, joint gating yields a MAP mean absolute error of 7.3 mmHg, compared with 9.2 mmHg for random selection at the same coverage. The gated agent achieves 60.0% agreement with recorded actions at 81.0% decision coverage, compared with 55.0% agreement and 99.5% coverage for an ungated ensemble agent. Component ablations show lower factual risk with joint gating than with either gate this http URL results characterize a reliability--coverage trade-off in selective clinical simulation; they do not identify individualized treatment effects or establish improved patient outcomes.
- [465] arXiv:2611.18459 [pdf, html, other]
-
Title: Callo: Serving Cache-Miss Requests with Speculative Decoding on Invalidated KV under Prefill-Decode DisaggregationComments: 24 pagesSubjects: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
Agent sessions may involve frequent context edits, which effectively invalidate cached KV state and force requests to wait for re-prefill before decoding. We present Callo, a disaggregated serving system that leverages invalidated KV for speculative drafts to reduce this latency, while verifying drafts against the exact KV after prefill before committing tokens. Callo makes effective use of decode resources by scheduling speculations during periods of underutilization and preempting speculations when drafts are likely to not be accepted. We evaluated Callo on a two-GPU testbed using SWE-agent sessions across three model families, observing all-request speedups of 1.05-1.19x over vLLM at moderate load. We demonstrate that these gains are also realized at scale through cluster-level simulations driven by real-world traces.
- [466] arXiv:2611.18462 [pdf, html, other]
-
Title: Sample-efficient prediction of high-entropy alloy strength through outcome-aware robust sequential learningSubjects: Machine Learning (cs.LG)
Accurate prediction of high-entropy alloy (HEA) strength can guide composition and processing decisions; however, obtaining tensile-test data for HEAs can be costly and experimentally demanding. This study proposes Robust Outcome-aware Sequential Experimentation for High-Entropy Alloys (ROSE-HEA), a sample-efficient framework for predicting ultimate tensile strength (UTS) from alloy composition and processing conditions. Its acquisition policy anticipates how possible UTS outcomes update predictions and the support for complementary Gaussian process (GP) descriptions. It then reassesses the worst-case expected predictive risk remaining across a reference domain to select informative candidates. A retrospective evaluation on 501 HEA records compares ROSE-HEA with expected-improvement (EI) and upper-confidence-bound (UCB) Bayesian optimization (BO), random GP sampling, and eight conventional regressors across 30 matched runs with shared initial observations and test conditions. Starting from 25 initial observations and acquiring 100 additional responses, ROSE-HEA achieves the lowest mean root-mean-square error (RMSE), at 152.84 MPa, and the highest mean R^2, at 0.7443. Mean RMSE is reduced by 6.16% and 6.37% relative to EI and UCB, respectively. The selected observations improve prediction across six regressor families and provide broader design-space coverage. ROSE-HEA also surpasses the strongest conventional regressor's mean predictive performance using 125 rather than 185 observations, corresponding to 32.4% fewer training observations. These findings demonstrate the potential of outcome-aware robust sequential learning to improve UTS prediction in HEAs under constrained training budgets.
- [467] arXiv:2611.18463 [pdf, html, other]
-
Title: VellSSLH: Sound Type-Guided Selective Speculative Load Hardening for LLVM IRSubjects: Cryptography and Security (cs.CR)
Constant-time cryptographic programs written in C can still leak secrets during speculative execution. Mispredicted branches may allow instructions to access secret data and expose it through observable behavior. Speculative load hardening (SLH) addresses these leaks, but protecting values unnecessarily can introduce significant runtime overhead. We present VellSSLH, which selectively inserts speculative load hardening at LLVM IR. Programmers specify Public and Secret inputs in annotated C programs, and our analysis infers which values may become secret during speculation. The transformation then inserts misspeculation-flag operations and protections only where needed. VellSCT, our extracted checker, validates the resulting LLVM IR.
We formalize speculative constant-time (SCT) security by extending Vellvm's LLVM semantics with speculative execution, attacker directives, and leakage observations. Our security type system reasons about SSA joins, symbolic values, poison, pointer provenance, and observable faults, and supports declared declassifications and trusted external services. In Rocq, we prove that programs accepted by the checker satisfy SCT, and that a reference hardening transformation preserves architectural behavior. Evaluated optimized variants are rechecked for SCT but are not covered by this preservation proof. Our mechanized guarantees apply at LLVM IR; security of the generated binary additionally depends on backend and hardware assumptions.
We evaluate VellSSLH on 24 cryptographic entries on a bare-metal Intel processor. VellSSLH achieves 4.03% lower geometric-mean runtime than LLVM SLH, with 13 comparisons favoring VellSSLH, eight favoring LLVM SLH, and three inconclusive. For X25519, VellSSLH improves runtime by 22.69% over LLVM SLH with only 1.86% overhead over unprotected code. Compared with Ultimate SLH, VellSSLH achieves 19.65% lower runtime. - [468] arXiv:2611.18464 [pdf, html, other]
-
Title: Typing Matters: Muscle, Language and Typing Behaviour in Live Keystroke Decoding from sEMGSubjects: Artificial Intelligence (cs.AI)
Surface electromyography (sEMG) from the wrists can recover typed text without a keyboard. It is usually posed as a signal-inference problem: characters are read out of the muscle signal and character error rate is reported offline. We argue that this framing removes most of the problem. The decoder is constrained simultaneously by the muscles, by the language being typed, and by typing behaviour such as typos, corrections, backspaces and line breaks, and a usable system must predict keys live while the hand keeps moving. Existing systems fall short on two counts: (1) they process sEMG in fixed windows and carry nothing from one window to the next, discarding context accumulated over a session, and (2) they lack any understanding of the linguistic structure that drives typing, so every keystroke is inferred from the muscle signal alone. We propose a fully causal, zero-lookahead sEMG typing decoder built on a state space model (SSM), in which a single model decodes a whole session without windowed partitions, conditioned on what has already been typed. We achieve this by fusing a traditional typing decoder with an efficient autoregressive language model, instilling both sEMG and linguistic understanding in one model. To train such a model for per-character typing rather than whole-word text, we introduce a retokenization method that rewrites ordinary prose as human-like typing, with keystrokes, typos, backspaces and line breaks included, converting WikiText into a large typing corpus where none existed. Our model, having never seen a single sEMG-to-character transcript in typing, achieves a state-of-the-art character error rate of 25.13% on emg2qwerty's cross-user benchmark, a 5.4% relative improvement over the previous best with 6.7 times fewer parameters, while running strictly causally with no future context.
- [469] arXiv:2611.18467 [pdf, html, other]
-
Title: Score-based neural ordinary differential equations for reaction-diffusionsSubjects: Numerical Analysis (math.NA); Machine Learning (cs.LG)
A diffusion equation can be rewritten as a transport equation driven by the score function of the evolving density. Conventional probability-flow formulations, which move particles deterministically with this score-driven velocity, typically conserve total mass. In this work, we develop a Lagrangian method for reaction-diffusion equations in which each particle carries a position and a physical weight. The characteristic ODEs incorporate source terms into the evolution of the log-density and its gradient, while particle weights track changes in mass. We combine this representation with a learned velocity field and weighted particle training to approximate these dynamics. For non-increasing reaction rates, a relative-entropy inequality supports the mass-normalized self-consistency residual as the training objective. This objective can control the error of the induced density when the growth rate is bounded. The algorithm divides the time interval into short windows. In each window, a frozen velocity generates the particle data for a regression of the network velocity. We draw the particles from period-adapted initial sampling distributions and weight them in the loss partly by their predicted future mass, because the reaction can make regions with little initial mass important later. We numerically investigate the method on the Ornstein--Uhlenbeck (OU) equation with a linear source term, Fisher--KPP equations, and Wasserstein--Fisher--Rao gradient flows.
- [470] arXiv:2611.18468 [pdf, html, other]
-
Title: Disenfranchisement in Spatial Voting GamesSubjects: Computer Science and Game Theory (cs.GT); Theoretical Economics (econ.TH)
The questions of voting rights, group membership, and the boundaries of acceptable discourse all involve a recursive, self-stability issue, wherein the voters, group members, and acceptable commentators are often the ultimate arbiters of their own rules for inclusion and exclusion. In order to understand how ideological or issue diversity may impact the stability of such inclusion and exclusion, we introduce franchise elections, where voting members of a group vote to change the electorate. In particular, when agents are positioned inside an ideological or issue space, a social choice function that aggregates agent viewpoints induces a preference for agents over group composition, where agents prefer groups that improve the social choice function outcome for them. We demonstrate that when ideological space is two-dimensional, these franchise elections can generate complex behavior and that this complexity exists for any social choice function. Using the Yolk social choice function from voting theory, we illustrate that franchise elections can predict that agents attempting to introduce a new ideological axis perpendicular to an existing left-right spectrum risk disenfranchisement unless they ally with either side of the spectrum, and that franchise elections can cycle with any period.
- [471] arXiv:2611.18470 [pdf, html, other]
-
Title: Stochastic Games with Distributional Entropy ObjectivesSubjects: Computer Science and Game Theory (cs.GT); Formal Languages and Automata Theory (cs.FL)
We introduce \emph{stochastic games with distributional entropy objectives}, two-player zero-sum games whose configurations are probability distributions over the states of a finite arena and whose stage reward is Shannon entropy. Over an infinite discounted horizon, MIN seeks to concentrate the distribution, while MAX seeks to disperse it. Because entropy is nonlinear, it matters whether strategic and transition randomness are averaged before entropy is evaluated or sampled first. We formalise three resulting semantics: averaged-distribution, sampled-profile, and sampled-transition.
This semantic choice fundamentally changes the game-theoretic structure. Averaged-distribution games need not be determined, even on a four-state deterministic arena, whereas both sampled semantics are determined and admit optimal stationary randomised strategies. We give contractive Bellman characterisations and convergent finite-horizon approximations for the upper and lower averaged values and the two sampled values. Moreover, even the lower value under averaged-distribution semantics dominates the value under either sampled semantics, while the sampled-profile and sampled-transition values are incomparable in general.
For fixed rational discount and positive rational promise gap, the averaged-distribution (upper and lower value) gap problems and both sampled gap problems lie in PSPACE; for a single entropy minimiser, the averaged and sampled-profile problems lie in NP. Finally, under all three semantics, entropy optimisation is NP-hard to approximate within a suitable fixed additive constant, already for deterministic one-player games, where pure memoryless strategies are optimal. - [472] arXiv:2611.18471 [pdf, html, other]
-
Title: Do Verifier Judgments Survive Abstraction? Measuring Verification Transport for LLM Agent EvaluationComments: Accepted at NeurIPS 2026 Workshop on Who Verifies the Agents? Toward Reliable Agent DevelopmentSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
LLM-based verifiers are increasingly used to evaluate agent executions, but their reliability is typically measured under a single representation of the trajectory. In deployed agent systems, the same execution may be exposed through composed tools, delegation boundaries, or compacted histories, changing what execution evidence is directly observable to the verifier. It therefore remains unclear whether correct verifier judgments persist across such abstraction boundaries. We introduce Verification Transport, a reliability test that measures judgment stability under abstraction-induced changes in execution observability. We construct matched native, abstracted, and restored views of the same frozen executions while holding the task, realized trajectory, outcome, ground truth, evaluation criterion, and verifier fixed. Across four agent-evaluation benchmarks, seven verifier backbones, four verification protocols, and three abstraction boundaries, we find substantial transport failure, reaching about 31% under macro abstraction and delegation. Verification protocols exhibit distinct failure modes: direct verification more often produces incorrect verdicts, whereas evidence-grounded and deliberative protocols tend to abstain under limited observability. This suggests that lower transport failure rates do not necessarily imply greater reliability. Restoring the corresponding execution provenance recovers 85.5% of transport failures, while recovery is lower when provenance is lossy or unavailable. These results show that native-view accuracy alone is insufficient to characterize verifier reliability when execution evidence crosses abstraction boundaries.
- [473] arXiv:2611.18476 [pdf, html, other]
-
Title: Beyond Blossom: Accelerating MWPM Decoding for Surface Codes With Local Predecoding and Architecture-Aware OptimizationsComments: 11 pages, 8 figuresSubjects: Information Theory (cs.IT); Performance (cs.PF); Quantum Physics (quant-ph)
Minimum-weight perfect matching (MWPM) is widely used to decode surface codes, but its computational cost increases with code distance. We present Beyond Blossom, a state-of-the-art CPU-based MWPM decoder for surface codes. It combines algorithmic optimization through local predecoding, architecture-aware optimization, and pipeline-wide parallelism to reduce matching workload and execution overhead while preserving MWPM optimality. Across rotated surface codes with distances 9-33 and circuit-level physical error rates of 0.05%-0.3%, on a single CPU core, Beyond Blossom achieves a 4.23$\times$ geometric-mean speedup and up to 5.58$\times$ speedup over Sparse Blossom in PyMatching. At distance 33 and a physical error rate of 0.1%, with 8 CPU cores, decoding a block of 33 measurement rounds takes 21.40 $\mu$s on average, equivalent to 0.65 $\mu$s per round.
- [474] arXiv:2611.18477 [pdf, html, other]
-
Title: Evaluating Feedback-Driven Black-Box Processes via Stochastic Modeling and SimulationSubjects: Software Engineering (cs.SE); Machine Learning (cs.LG)
We study feedback-driven black-box processes across software engineering and machine learning. These processes are inherently sequential: policies generate inputs, observe outputs/feedback from systems, and use the feedback to guide the design of subsequent inputs. While many policies are developed under structural assumptions on the underlying systems, many real-world systems operate without such knowledge, necessitating black-box policies that lack a unified evaluation framework. We introduce a stochastic framework that models and evaluates such policies using Markov chains and simulation, thereby connecting empirical practice with theory. We instantiate the framework in mutation-based fuzzing (widely used for discovering software bugs via sequential input mutation), by a novel Markov chain model of fuzzing whose transitions are governed by the mutation strategy and whose feedback is given by program execution. We develop a rigorous evaluation paradigm with new performance measures, sound estimators, and evaluation procedures with provable guarantees. Guided by the theory, we evaluate 9 fuzzers on 18 widely used fuzzing benchmarks and conduct case studies in 4 non-fuzzing domains. These results show that the framework provides a principled basis for evaluation while connecting to practice, with implications beyond fuzzing.
- [475] arXiv:2611.18482 [pdf, html, other]
-
Title: MMFlash: Faster Speculative Decoding for Vision-Language ModelsComments: Work in process, website: \url{this https URL}Subjects: Computer Vision and Pattern Recognition (cs.CV)
Speculative decoding accelerates generation by letting a lightweight drafter propose tokens for parallel verification by a target model. Diffusion drafters such as DFlash propose an entire token block in a single forward pass, but mainly focus on text-only tasks, leaving multimodal drafting underexplored. To study this setting, we first introduce a metric of token-level visual dependence based on the KL divergence between next-token distributions with and without visual input. We find that acceptance rates decrease from 46% in the lowest visual-dependence quartile to 22% in the highest, with this disparity persisting after multimodal training. Motivated by this observation, we propose MMFlash, a simple and effective framework that adaptively upweights visually dependent tokens that the drafter predicts poorly. For long video inputs, MMFlash further employs sparse attention to reduce redundant attention computation and improve drafting quality. Across various multimodal benchmarks, MMFlash achieves lossless decoding speedups of up to $3.58\times$, outperforming EAGLE-3 and DFlash by up to 195.4% and 17.1%, respectively.
- [476] arXiv:2611.18484 [pdf, other]
-
Title: From AI-Enabled Remote Patient Monitoring Toward Agentic Robotics: Embodiment, Governance, and Care IntegrationComments: 34 pages, 6 figures, 3 tables. Accepted for publication in Annual Review of Control, Robotics, and Autonomous Systems, Volume 10 (2027)Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Software Engineering (cs.SE); Systems and Control (eess.SY)
AI-enabled remote patient monitoring is evolving beyond passive data collection toward agentic AI systems that interpret patient states, coordinate care, and initiate action across domestic and clinical settings. This structured scoping review finds that mature agentic care has not yet emerged: Most systems remain human led, while stronger decision-action coupling and embodied AI remain largely confined to prototypes and pilots. Healthcare robotics makes the transition from monitoring to action especially visible by translating AI-generated recommendations and decisions into speech, movement, prompting, physical assistance, escalation, and care coordination, making embodiment, clinical integration, human oversight, and governance fundamental design requirements. Further progress requires safe and interoperable care architectures in which intelligent components operate through clinically validated and accountable workflows. We propose agentic robotics as a future direction for supervised systems embedded in the care ecosystem, connecting AI agents and robots with clinicians, caregivers, and patients through auditable workflows that can be interrupted and remain subject to human oversight and explicit AI governance.
- [477] arXiv:2611.18485 [pdf, html, other]
-
Title: Interactive Program Induction with Language Model PriorsComments: Preprint. 28 pages, 5 figures, 9 tablesSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Program induction refers to the process in which a learner aims to recover a program from a few given input-output examples. This paradigm broadens programming access by letting users demonstrate the rule-based behavior they want to apply to unseen data, instead of having to implement it themselves. However, recovering a program from finite input-output examples is fundamentally ambiguous, as a finite set of input-output examples is theoretically consistent with infinitely many programs whose output diverges on unseen inputs. We view this problem from a Bayesian perspective, where, given a prior $p^0$ over programs, identifying the target program means concentrating posterior probability mass on it by observing additional input-output examples. In the interactive program induction setting, the learner chooses maximally informative inputs and queries an oracle, such as the user, for their outputs. We derive a greedy query-selection algorithm and show that it terminates after $\mathcal{O}(\frac{1}{\gamma}\log 1/p^0(f^\star))$ informative oracle queries, and returns the target program $f^\star$ up to approximate equivalence when no other class outweighs it under the prior. The algorithm uses a language model as the prior under a PAC-style relaxation of program equivalence, where the query-selection objective is estimated using importance sampling. On the CodeARC benchmark, the algorithm improves accuracy over the LM only agent by up to 22 percentage points while observing up to 10.5 fewer input-output examples.
- [478] arXiv:2611.18491 [pdf, html, other]
-
Title: When Will You Hand It Over? A Hidden Start Delay Makes Self-Play Handover Robust to Unfamiliar PartnersComments: CoRL 2026, Human-Centered Robot Learning and Interaction (HumanRLI) Workshop. Code: this https URLSubjects: Robotics (cs.RO)
Cooperative self-play can teach a humanoid to receive an object without any human data, but the partner it meets after training is never the one it trained with; we show that a fluent self-play handover policy breaks with unfamiliar partners, identify the main cause, and repair it. In a physics-based parcel handover between a Unitree G1 and a learned giver, both trained by alternating PPO self-play, the receiver completes 95% and 90% of handovers with its own partner (two training seeds) but only 31% with scripted partners it has never met. Paired interventions isolate the cause: under a shared reward the learned giver always starts its offer at once, so the receiver learns a single rendezvous, and delaying its own partner by one second cuts success from 95% to 6% (87% to 34% in the second seed). We introduce hidden start-delay randomisation (HSD), a one-line rule on the learned partner that the receiver never observes. After 300 further updates, success with unseen scripted partners rises from 31% to 80% in both seeds, against 35--41% for a budget-matched control. The gain holds on a fresh scenario seed, on eleven held-out partner conditions and on partners that start later than any training delay; it cuts high-force contact three- to five-fold and needs no retraining of the partner.
- [479] arXiv:2611.18493 [pdf, html, other]
-
Title: TradeGauntlet: Evaluating Compliance Controls for Autonomous Trading AgentsComments: 21 pages, 5 figures. Code and saved results: this https URLSubjects: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE)
Autonomous trading agents need controls that prevent problematic orders without making legitimate tasks impractical. We study this aspect of agentic compliance with TradeGauntlet, a replayable limit-order-book simulator that separates detection, prevention, and execution cost. In 1,680 simulated sessions with Claude Haiku 5.5 and Sonnet 5.5, all conduct alerts were self-matches. Gates eliminated those alerts by rejecting self-crossing orders, while the aggregate-gate condition reduced Sonnet's two-sided fills by about 15%. A conduct reminder reduced Sonnet's alert rate but reduced Haiku's fills without a detectable alert benefit. These results show why zero alerts alone do not establish an agent's compliance. Scripted probes and two search methods test weaknesses beyond those observed in the LLM runs. The searches find split-order schedules that evade the implemented conduct rules; a learned baseline catches all 2,400 tested searched episodes but also flags 15% of a large maker's sessions. Aggregate caps reduce the response-dependent gain unevenly, with smaller reductions when only near-touch depth affects prices. A response fitted to five Nasdaq stocks and a separate ABIDES experiment expose the dependence of these estimates on market assumptions. A distinct adversarial prompt elicited refusals from Sonnet in all 240 rounds, yielding no candidate strategies to evaluate. TradeGauntlet provides reproducible tests of controls around trading agents; the experiments do not establish legal compliance or real-market detection accuracy.
- [480] arXiv:2611.18494 [pdf, html, other]
-
Title: ARCHER: From Rationale Imitation to Executable Reasoning in Small Language ModelsSubjects: Artificial Intelligence (cs.AI)
Small language models are attractive for local, low-cost, and privacy-sensitive deployment, but they remain unreliable on mathematical reasoning tasks that require decomposition, exact computation, tool use, and verification. We introduce \textbf{ARCHER}: \textbf{A}ction-trace \textbf{R}einforcement for \textbf{C}alculator-grounded \textbf{H}ierarchical \textbf{E}xecutable \textbf{R}easoning, a two-stage framework for training small language models (SLMs) to produce executable and auditable reasoning trajectories. In the first stage, ARCHER compiles free-form chain-of-thought rationales into typed action traces containing reasoning steps, calculator calls, deterministic runtime observations, verification actions, and final answers. This shifts the supervision target from imitating explanatory text to learning a tool-grounded reasoning policy. In the second stage, ARCHER refines this policy through asymmetric hierarchical reinforcement learning, using grouped and conditionally gated trajectory rewards for answer correctness, tool grounding, process discipline, and trace alignment. Across five standard sub-10B backbones, ARCHER improves over the strongest tool-augmented baseline by 5.51 percentage points on macro-average across benchmark accuracies, including gains of 8.95 points on MATH-500 and 5.33 and 5.34 points on AIME-2024 and AIME-2025. Evaluations on two additional reasoning-specialized 7B backbones further show that executable grounding remains beneficial after reasoning-focused reinforcement learning or distillation. These results demonstrate that executable supervision and reward-guided behavioral refinement provide a general and effective training paradigm for compact mathematical reasoners.
- [481] arXiv:2611.18496 [pdf, html, other]
-
Title: Toward Resilient Self-Driving Laboratories through Telerobotic InterventionsZiheng Wu, Adrian Vecina Tercero, Andre A. Fatehi, Abhik Sarkar, Brian Giera, Eric B. Duoss, Christopher M. Spadaccini, Nikhil Deshpande, Aldair E. GongoraSubjects: Robotics (cs.RO)
Self-driving laboratories (SDLs) accelerate scientific research by integrating robotics, automated experimentation, and machine learning to execute experiments autonomously. Biology and chemistry have adopted these systems. However, increasing SDL complexity heightens susceptibility to unexpected faults in automated workflows, highlighting the need for efficient recovery mechanisms. In this study, we investigate telerobotic interventions to enhance SDL resilience against unexpected failures. We developed a teleoperation interface enabling human operators to resolve disruptions during automated tasks. The interface enables precise control of a six-degree-of-freedom (DOF) robotic arm (UR3e) using a 3D mouse (3Dconnexion SpaceMouse Pro). Two velocity-based controller strategies were evaluated: (i) a coupled mode mapping all six DOFs to the 3D mouse's central knob and (ii) a decoupled mode separating translational and rotational motions. To evaluate the framework, teleoperators intervened across three SDL fault scenarios. In the first study, comparative trials showed that the decoupled controller was associated with shorter completion times and reduced rotational motion. In a second study, teleoperators performed an aiming task with and without reference-path guidance. Users were more target-driven and risk-tolerant without guidance but showed behavioral differences under different guidance strategies. These findings suggest that controller configuration and reference-path design influence operator performance during SDL fault recovery, informing teleoperation interface optimization. This work demonstrates the potential feasibility of teleoperation as a human-in-the-loop recovery mechanism for SDL workflows, combining robotic precision with human adaptability and problem-solving. These findings also provide a foundation for future studies in shared autonomy and semi-autonomous intervention.
- [482] arXiv:2611.18498 [pdf, html, other]
-
Title: A Common Improved Bound for Asymmetric TSP and Path TSPSubjects: Data Structures and Algorithms (cs.DS)
We give polynomial-time $(R_*+\varepsilon)$-approximation algorithms for asymmetric TSP and asymmetric path TSP, for every fixed $\varepsilon>0$, where $R_*=6+\sqrt2+\sqrt{15+6\sqrt2}\approx12.2604$. Both guarantees hold relative to the corresponding subtour relaxations, whose integrality ratios are therefore at most $R_*$. This improves the previous guarantees of below $14.7$ for asymmetric TSP and $15$ for asymmetric path TSP [Vygen, 2026]. Our algorithms strengthen subtour covers and recursive augmentation in the laminar framework [Svensson, Tarnawski, Vegh, JACM'20; Traub, Vygen, 2025]. The recursive analysis retains the paths through contracted regions and accounts jointly for their cost and the dual mass they visit. We also improve the reduction from asymmetric path TSP to vertebrate pairs [Traub, Vygen, 2025] by comparing two initial walks with complementary completion bounds. Applied to Vygen's existing vertebrate-pair algorithm, this reduction alone improves the path factor from $15$ to below $14.7$, matching his asymmetric TSP guarantee.
- [483] arXiv:2611.18501 [pdf, html, other]
-
Title: In Situ Monitoring and Predictive Modeling of Harmful Algal Blooms with Autonomous Surface VehiclesNenyi Kweku Nkensen Dadson, Jean Paul Gonzalez, Daniel P. Arabie, Ryan J. Cedel, Dorin Boldor, Corina BarbalataSubjects: Robotics (cs.RO)
Harmful Algal Blooms (HABs) pose significant environmental, economic, and public-health challenges in coastal and inland water bodies, creating a need for scalable methods to collect spatially distributed water-quality data. This paper presents an autonomous surface vessel-based sensing system for in-situ water-quality monitoring related to harmful algal bloom detection. The system integrates a BlueBoat Autonomous Surface Vessel (ASV), a motorized winch for depth-controlled sonde deployment, a multiparameter water-quality sonde, side-scan sonar, and an event-camera payload. Autonomous waypoint missions were used to collect measurements across selected sampling regions and multiple depths, enabling both horizontal and vertical characterization of environmental conditions. Field deployments were conducted in Lake Pontchartrain, Louisiana, with additional measurements collected in False River, Louisiana to increase dataset diversity. The collected measurements included algal pigment indicators, nutrient concentration, turbidity, temperature, and other water-quality parameters relevant to bloom formation. In addition, complementary sonar and event-camera data were collected during vehicle operation to provide contextual environmental information. A machine-learning classification approach was also evaluated using the collected water-quality measurements to support HAB and non-HAB classification. The results demonstrate the feasibility of using a modular ASV platform for repeatable, multi-modal water-quality data collection and support future development of autonomous monitoring systems.
- [484] arXiv:2611.18502 [pdf, html, other]
-
Title: EPIGRIDS: Documentation and Validation of Seven Open Synthetic Transmission Grid CasesJonathan M. Snodgrass, Scott Greene, Thomas J. Overbye, Christopher L. DeMarco, Bernard C. LesieutreComments: 10 pages, 2 figures, 6 tables. Submitted to the IEEE Open Access Journal of Power and Energy. Cases: this https URL load series: this https URL construction method: arXiv:2611.16222Subjects: Systems and Control (eess.SY); Optimization and Control (math.OC)
This paper documents and checks EPIGRIDS, seven open synthetic transmission grid cases from four independently built networks, made with an optimization-based construction method instead of the graph-statistics method behind most large public synthetic grids. The cases range from 250 to 78,484 buses, with geolocated substations, voltages up to 765 kV, and line parameters taken from conductor data. Each ships under an open license in three common power flow formats that agree at the solved operating point, and the six larger cases have a companion hourly, bus-level load series for one year. The Texas, Florida, and Michigan, Ohio, and Indiana cases are sub-networks of the Eastern case. Every case solves an alternating-current power flow from its stored operating point. With parallel circuits merged, the Eastern and Western cases are 9 and 16% denser than the real interconnections, while graph-statistics cases on the same footprints lie within 6% and meet more published realism criteria. A full single-outage contingency analysis completes on every case, and in an optimal power flow benchmark every case except the New England base case and one of its 12 operating points solves in polar form.
- [485] arXiv:2611.18505 [pdf, html, other]
-
Title: Disturbance-Aware Informative Planning in Spatiotemporal EnvironmentsComments: Submitted to Learning for Dynamics and Control Conference 2027. Project website: this https URLSubjects: Robotics (cs.RO)
Autonomous robots are widely used for persistent monitoring of spatiotemporal environments, where quantities of interest (QoI) change across space and time and require repeated measurements. In practice, robot motion can be affected by external disturbances that vary across the environment. When such external disturbances are not considered in trajectory optimization, it can lead to inaccurate predictions of robot motion and, consequently, the information acquired about the QoI during execution. However, estimating disturbances throughout the environment can be inefficient, as disturbances in regions with little overlap with monitoring trajectories have limited impact on monitoring performance. We therefore propose a two-stage informative planning framework for unknown, spatially varying but time-invariant disturbances, combining informed disturbance sampling with disturbance-aware trajectory optimization. First, we identify where QoI measurements are needed and generate alternative paths to these locations using the environment geometry. We then prioritize disturbance measurements based on their expected reduction of disturbance uncertainty along these paths. Second, rather than optimizing information acquisition based on nominal robot motion, we incorporate the learned disturbance belief into trajectory optimization to account for its effects on robot motion and information acquisition. Across four simulation scenarios, the proposed method reduces disturbance-learning time by 25.0--52.3 percent compared with environment-wide estimation and identifies paths that appear costly under nominal dynamics but become preferable when disturbances are considered.
- [486] arXiv:2611.18506 [pdf, html, other]
-
Title: Agentic induction of common data elements and information extraction from sarcoma case report abstractsDaniel Reichenpfader (1 and 2), Kwaku Appiagyei (3), Manovah Nadarajah (4), Ahmed Reda Bahr (5), Irina Filchenko (6), Fabio Dennstädt (7, 8 and 9) ((1) Institute for Patient-Centered Digital Health, Bern University of Applied Sciences, Bern, Switzerland, (2) Faculty of Medicine, University of Geneva, Geneva, Switzerland, (3) F. Hoffmann-La Roche AG, Basel, Switzerland, (4) Kantonsspital Aarau, Aarau, Switzerland, (5) University of Alexandria, Alexandria, Egypt, (6) Department of Neurology, Inselspital, Bern University Hospital and University of Bern, Bern, Switzerland, (7) Data Semantics and Decision Making Institute (DSDMI), St. Gallen, Switzerland, (8) Department of Radiation Oncology, Inselspital, Bern University Hospital and University of Bern, Bern, Switzerland, (9) Department of Digital Medicine, Inselspital, Bern University Hospital and University of Bern, Bern, Switzerland)Subjects: Computation and Language (cs.CL)
Common data elements (CDEs) support clinical data comparability but are usually defined manually. We investigated whether large language model agents could derive a CDE schema from sarcoma case report abstracts and apply it at scale. An agent proposed 72,909 candidate CDEs from 24,588 PubMed abstracts; agentic consolidation yielded 1,570 CDEs. We retained 53 supported by at least 50 abstracts and attempted evidence-linked extraction across the corpus. Processing succeeded for all retained CDEs in 79.2% of abstracts, including responses indicating non-mention. Two human reviewers approved 83.0% of retained CDEs for relevance, 67.9% for clarity, and 56.6% for correctness. In 100 sampled abstracts, exact agreement with the value extracted by the LLM was 94.9% among 1,116 pairs with a non-empty answer (abstract-level bootstrap 95% CI, 93.0-96.4%). Errors concentrated in anatomical and temporal fields. These findings support agent-assisted schema drafting followed by expert curation before large-scale extraction.
- [487] arXiv:2611.18507 [pdf, html, other]
-
Title: Asymptotic Margins and Feature Confinement in Conic Classification Heads under Softmax Cross-EntropyComments: 22 pages, 2 figures. Companion paper: to be uploaded. Code: this https URLSubjects: Machine Learning (cs.LG); Signal Processing (eess.SP)
Classification heads built on conic functions give each class a bounded region of the feature space, and such regions are used to reject samples of unseen classes. We ask whether these regions also keep the features bounded when the head is trained with softmax cross-entropy. For a head whose class scores combine a linear term with a Euclidean norm, we first show that the region of each class is bounded by a quadric of revolution, whose eccentricity is the reciprocal of one learned ratio. We then attach to each class an asymptotic margin, which compares the growth of its score with the growth of all the other scores and can be computed by a second-order cone program. The loss of a class has infimum zero exactly when its margin is positive, and it then approaches zero only along diverging features. When the margin is negative, the loss is coercive and its minimizers lie in an explicit ball. Since softmax depends only on the differences between scores, a class with a bounded region can still have a positive margin, so bounded regions do not confine the features. The result extends to other norms, including the polyhedral conic functions. On CIFAR-10 and CIFAR-100, all 1210 cells of 67 trained conic heads have positive margins, whereas a prototype head keeps nine of its ten cells bounded, as the theory predicts.
- [488] arXiv:2611.18511 [pdf, html, other]
-
Title: On numerical issues of divergences between nearly equal distributionsSubjects: Machine Learning (cs.LG); Numerical Analysis (math.NA); Machine Learning (stat.ML)
Barron et al. (2026) recently identified numerical issues in the softmax and soft cross-entropy loss. We show that similar issues are present in standard divergence operations between categorical distributions. This includes the forward and reverse Kullback-Leibler (KL) divergence and generalized Jensen-Shannon divergence (JSD), as well as Monte Carlo KL estimates and entropy. When the distributions are nearly equal, rounding and cancellation issues affect the forward and gradient computations. These issues can affect settings using divergences in optimization objectives, such as knowledge distillation, reinforcement learning, and proximal methods. We propose numerically stable reformulations and provide example implementations for each setting, validated by property-based tests.
- [489] arXiv:2611.18513 [pdf, html, other]
-
Title: MeReader: An Offline, Privacy-Preserving, Progress-Aware AI Assistant for Narrative eBook ReadingComments: 15 pages, 4 figures. Preprint (submitted version) of a paper published in ICCCI 2026, CCIS vol. 3044, SpringerJournal-ref: Advances in Computational Collective Intelligence (ICCCI 2026), CCIS vol. 3044, pp. 321-335, Springer, 2027Subjects: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
This paper introduces MeReader, a privacy-preserving, offline and progress-aware AI assistant designed to function as a contextual companion for narrative reading on consumer hardware. Unlike conventional Large Language Model (LLM)-based systems that process documents/books online in an all-texts-at-once type of way, thereby risking narrative spoilers, MeReader confines both retrieval and generation to the reader's distinct position in the text. The system integrates a hybrid retrieval architecture, combining vector similarity, BM25, and summary embeddings, to ensure high fidelity to the reader's current context. We evaluate the system across classical English books using both quantitative comprehension metrics and qualitative LLM-based independent judging. The results identify a quality-latency frontier, showing that mid-sized local models (2b-4b) can achieve comprehension parity with larger counterparts (5b+) while maintaining acceptable response quality. Importantly, the proposed progress-aware boundary mechanism is shown to prevent future content leakage without compromising retrieval quality and correctness. These findings suggest that a localized, privacy-centric approach can augment, rather than replace, the digital reading experience, offering a good alternative to cloud-dependent services. Therein, there remains a lack of amalgamate systems which combine progress-aware, privacy-preserving, and narrative eBook reading with AI support. It is in this very gap that the development of MeReader came to be; synthesizing said strands into an offline, book-constrained assistant focused on enhancing comprehension without compromising narrative continuity or user privacy.
- [490] arXiv:2611.18514 [pdf, html, other]
-
Title: Can Mechanistic Interpretability Help Explain LLM Agent Failures? A Case Study on Web Element HallucinationSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Mechanistic Interpretability (MI) has made significant progress in explaining the internal computations of Large Language Models (LLMs); however, these advances have yet to demonstrate comparable success in explaining the decision-making processes of LLM-based agents. This paper examines the challenges MI tools face when applied to realistic agentic failures. We focus on one such failure: web agents hallucinating actions that are not available on a website. We apply a series of mechanistic analyzes to Qwen2.5-7B-Instruct as it interacts with live websites. We find that a linear probe trained on the last prompt token can predict hallucinations before generation begins (0.80 AUC) and steering along a probe direction corrects 85\% of hallucinations, but disrupts 7\% of valid actions. However, these results only provide a partial explanation of why the agent hallucinates. The probe direction is sensitive to prompt formatting, losing much of its steering effect after a minor change to how actions are represented. Additionally, circuit identification can be difficult as the computation is distributed across multiple components and positions, and the information the agent relies on varies across examples. We discuss how these observations challenge the assumptions underlying standard MI methods for circuit discovery and feature identification, and outline future directions for making these methods better suited to explain failures in LLM-based agents.
- [491] arXiv:2611.18516 [pdf, html, other]
-
Title: StreamEmbed: Real-Time Indexing from Pixels for Screenshot RetrievalSubjects: Computer Vision and Pattern Recognition (cs.CV); Information Retrieval (cs.IR); Machine Learning (cs.LG)
Screen history can serve as searchable visual memory for people and agents, but unlike an offline corpus, it must be encoded and indexed as each screenshot arrives. OCR-based indexing is fast but loses visual information; visual encoding takes over two seconds. Yet screens change incrementally: most pixels repeat within and across screenshots. This raises a question: do visual encoders need to recompute their representations from scratch for every new screenshot? We present StreamEmbed, a system for incremental visual embedding that exploits this redundancy. Guided solely by pixel differences, it combines spatial token pruning, which skips blocks that repeat a neighbor, with temporal KV reuse, which reuses cached keys and values for blocks unchanged across screenshots. A small learned adapter compensates for pruning and reuse to preserve fidelity to full-screenshot embeddings. On ScreenMemEval, our benchmark of 10,426 questions over 184 hours of screen histories from two humans and a computer-use agent, StreamEmbed sets a new accuracy-latency Pareto frontier on an M2 Pro GPU: it encodes up to 3.7x faster than full-screenshot encoding without sacrificing accuracy and improves accuracy over OCR-based indexing by up to 6.3 Recall@10 points at matched latency. A memory-efficient vector index stores a simulated year of screenshots in 0.13 GB of RAM. Together, these results make real-time indexing of screen history in its native visual form practical, laying a foundation for fully local visual assistants.
- [492] arXiv:2611.18517 [pdf, html, other]
-
Title: OpenLayout: An Open and Agentic-Driven Platform for Holistic VLSI Physical Design and BenchmarkingComments: Accepted by ICCAD2026Subjects: Hardware Architecture (cs.AR)
Physical design automation of very large-scale integrated circuits (VLSI) has made significant strides, progressing from single-stage placement and routing toward increasingly integrated full-flow implementation. Yet, existing approaches struggle with the vast design space optimization (DSO) arising from diverse technology nodes, circuit designs, and varied constraints. The substantial runtime of physical-design tools further exacerbates this challenge, leaving industrial workflows heavily dependent on experienced engineers to navigate the design space and achieve power, performance, and area (PPA) targets, incurring significant engineering effort.
We argue that large language models (LLMs) with strong reasoning capabilities, enriched with distilled human design expertise, offer a promising path toward efficient and automated DSO. Unlike traditional algorithms, LLMs can interpret complex design contexts and naturally incorporate human expertise expressed in natural language for reasoning, acting, observing, and adapting throughout the design process, a capability that conventional optimization methods cannot exploit. To this end, we present OpenLayout, an agentic platform that leverages multi-agent collaboration with embedded design knowledge to enable self-iterative DSO. We further introduce OpenLayout-Bench, a benchmark targeting advanced technology nodes with commercial EDA tools to systematically evaluate physical design automation capabilities of LLMs. Experimental results demonstrate that OpenLayout achieves state-of-the-art performance, with approximately 20% accuracy improvement over baseline methods. Our work is open-sourced at this https URL - [493] arXiv:2611.18519 [pdf, html, other]
-
Title: Box-conditional Hypothesis Testing for Out-of-Distribution Object DetectionComments: 18 pages, 7 figures, 13 tablesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Detecting out-of-distribution (OOD) objects is essential for ensuring the reliability and safety of object detectors in real-world deployment. We treat the decision on a detected box as a hypothesis test with the null hypothesis that the box covers an object of an in-distribution (ID) class. The null distribution changes with the size of the box and its contact with the image border, and the detector observes both. The test can therefore rank a box only against ID boxes of the same size and border contact, which gives a distribution-free bound on the false-alarm rate for each of them and uncovers the separation that one null distribution hides. The power of the test also rests on the layer it reads. Layers built only for the ID classes have no reason to keep what sets an unknown object apart, so the test reads the features before them. We propose \textbf{PLUM}, a training-free score that pools the last backbone stage inside the box and ranks its distance to the ID classes among ID boxes of the same size and border contact. On the standard benchmark, PLUM reaches the highest AUROC among the compared methods without external models, with 91.5 and 94.5 on PASCAL-VOC and 95.1 and 97.5 on BDD100K. Code: this https URL
- [494] arXiv:2611.18522 [pdf, html, other]
-
Title: Patient-specific Modeling of the Tricuspid Valve from 3D Transesophageal EchocardiographyDavide Tondi (1), Vijay K. Dubey (2), Collin E. Haese (2), Colton J. Kostelnik (3), Mrudang Mathur (4), Sumedh Seetharam (2), Amin Silvani (5), Alison M. Pouch (5), Jakob Christoph Voran (6 and 7), Felix Kreidel (8), Marnie Goldmann (8), Adrian Buganza Tepole (9), William D. Meador (3), Marcin Malinowski (10), Magda Piekarska (10), Austin Goodyke (10), Issam D. Moussa (11), Matthew R. Summers (12), Robyn Fong (4), Matthew M. Duda (4), Kelly C. Higa (4), William Hiesinger (4), Tomasz A. Timek (10), Jan N. Fuhg (13 and 14), Emiliano Votta (1), Manuel K. Rausch (2, 3, 13 and 14) ((1) Department of Electronics, Information and Bioengineering, Politecnico di Milano, Milan, Italy, (2) Walker Department of Mechanical Engineering, The University of Texas at Austin, Austin, USA, (3) Department of Biomedical Engineering, The University of Texas at Austin, Austin, USA, (4) Department of Cardiothoracic Surgery, Stanford University, Palo Alto, USA, (5) Department of Radiology and Bioengineering, University of Pennsylvania, Philadelphia, USA, (6) Department of Internal Medicine III, Cardiology and Critical Care, University Hospital Schleswig-Holstein, Kiel, Germany, (7) German Centre for Cardiovascular Research (DZHK), Partner Site North, Germany, (8) Department of Cardiology, Asklepios Hospital, Hamburg, Germany, (9) Department of Mechanical Engineering, Columbia University, New York, USA, (10) Division of Cardiothoracic Surgery, Corewell Health, Grand Rapids, USA, (11) Carle Illinois College of Medicine, University of Illinois at Urbana-Champaign, Urbana, USA, (12) Sentara Heart Valve Center, Sentara Heart Hospital, Norfolk, USA, (13) Department of Aerospace Engineering and Engineering Mechanics, The University of Texas at Austin, Austin, USA, (14) Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin, Austin, USA)Comments: Draft manuscript of 3D TEE to FEM modelling pipeline for human tricuspid valvesSubjects: Computational Engineering, Finance, and Science (cs.CE)
Purpose: Image-based patient-specific simulations of the tricuspid valve (TV) could support repair planning. However, clinical images incompletely resolve the chordal anatomy and cannot directly determine tissue properties, which hinders the development of such models. We developed and evaluated a framework that simulates the TV from clinical images and addresses these limitations using measurements on explanted human valves. Methods: We reconstructed ten functionally competent human TVs from 3D transesophageal echocardiography. Following manual segmentation, we automatically built finite element models with synthetic chordae and with leaflet thickness, chordal insertion density, and tissue properties informed by ex vivo human measurements. Next, we used shape matching and force-based chordal-length adjustment to simulate valve closure. We also assessed inter-observer segmentation variability and sensitivity to chordal insertion density, leaflet stiffness, leaflet thickness, and chordal size and material properties. Results: Per-valve median geometric errors were 0.61-1.13 mm relative to the morphed references and 0.55-0.95 mm relative to the image-derived surfaces. Coaptation covered 20-42% of the leaflet area, and median leaflet stresses ranged from 7.6 to 16.9 kPa. Inter-observer variability in segmentation most strongly affected coaptation area. Of the population-based inputs, insertion density and leaflet stiffness most affected geometric error. Thickness primarily affected stress, whereas chordal size and material properties had limited effects. Conclusion: Clinical echocardiography, supplemented with ex vivo human tissue data, enabled patient-specific simulation of ten human TVs. This framework provides a foundation for investigating valve mechanics and developing repair-planning tools, while sensitivity analyses identify priorities for model refinement and uncertainty assessment.
- [495] arXiv:2611.18523 [pdf, html, other]
-
Title: Mutable Inference: Architecting LLM Accelerators for Inference-Time LearningSubjects: Hardware Architecture (cs.AR)
Long-context LLMs must retain growing histories, but full attention makes each new token increasingly expensive. Inference-time learning (ITL) offers another approach. The model keeps training on the context during inference, retaining what it learns in its weights. Recent work shows that ITL can match full-context attention quality while attending only to a recent fixed-size window. This bounds attention costs but adds Learning (backward computation and weight updates) alongside Prediction (the forward pass) costs. Inference accelerators assume fixed weights, while training accelerators target complete training steps.
This work identifies two opportunities. Attention's non-matrix operations leave enough idle matrix capacity for Learning, while selected layers can tolerate delayed weight updates with little quality loss. We exploit both with Learning While Serving (LWS), an inference accelerator architecture that shares compute hardware between latency-critical Prediction and deadline-flexible Learning. Hardware-managed per-layer weight versioning allows Learning to publish updates while in-flight computations read older versions. Resumable reduction blocks let Learning kernels span fragmented idle windows and yield to Prediction, while dedicated softmax-gradient units relieve vector contention. On modeled datacenter and edge NPUs, LWS hides Learning within Prediction, sustaining Prediction-only throughput while preserving 99.3% of fully fresh Learning's quality benefit for TTT-E2E 3B at 128K tokens. Across nine models, LWS achieves 1.6x the throughput of serial execution and 1.3x that of the best prior co-execution architecture. RTL synthesis shows that LWS adds 1.5% to NPU core area without another MAC array. By efficiently turning idle inference capacity into learning capacity, LWS establishes inference-time learning as a first-class workload for future LLM accelerators. - [496] arXiv:2611.18524 [pdf, html, other]
-
Title: EventSpikeSeg: Memory-Aware Mixed-Signal Architecture for Event-Camera Spiking SegmentationSubjects: Hardware Architecture (cs.AR); Computer Vision and Pattern Recognition (cs.CV); Neural and Evolutionary Computing (cs.NE)
Event-camera segmentation requires persistent neural state despite sparse input events, making memory organization a central constraint on embedded implementation. We present EventSpikeSeg, a mixed-signal architecture combining resident capacitor memory with shared analog processing in GF180MCU 180-nm CMOS. Its 25-population spiking encoder-decoder has 143,824 synaptic weights but requires 30,976,000 membrane states at 440 x 640 resolution. The mapping specifies chronological 5-ms model updates, digital event/address services and 64-state capacitor banks. Characterization covers 22 module families and 74 operating points, distinguishing transistor simulations from digital-library estimates. Nominal results include 41.711 microwatts biased neuron-core power, 3.338-pJ positive terminal energy for a 1.4-microsecond cell read, and 0.531-mV selected-state drift over a 5-ms extracted-tile hold. Commit and coupling errors are reported separately. The software checkpoint records 27.976% mean intersection-over-union on a held-out six-class DSEC split; this is not hardware accuracy. Component layouts and simulation evidence quantify local implementation costs and expose storage replication and service scheduling as scaling constraints. A detailed SPICE-calibrated behavioral forward simulation of one full-resolution DSEC frame achieves 24.26% mean intersection-over-union, versus 26.22% for the matched PyTorch reference, using a compact 5-fF bank variant and external count/voltage calibration. Physical full-network operation, system energy and mismatch robustness remain unverified.
- [497] arXiv:2611.18527 [pdf, html, other]
-
Title: WaferTetris: Efficient Freeform Mapping on Wafer-Scale AcceleratorsSubjects: Hardware Architecture (cs.AR); Computation and Language (cs.CL)
Wafer-scale accelerators offer massive compute capacity through hundreds of thousands of processing elements (PEs) interconnected across a single wafer, but realizing this potential requires efficient application mapping. Their scale makes programming these systems and mapping applications onto them challenging. Existing approaches rely on template-based mappings to maintain productivity, but these templates restrict the mapping design space. In particular, they impose fixed geometric layouts over the PE grid, often leaving compute and memory resources underutilized.
In this work, we present WaferTetris, a system that removes these geometric restrictions through freeform mapping. Its Tetris-like strategy flexibly packs computation across the wafer, improving PE utilization and usable memory capacity while preserving PE-task code. We develop a recurrence-based throughput model that captures branch-and-join dependencies to guide spatial parallelism, and an architecture model with per-color gates that enable existing FPGA place-and-route algorithms to enforce WSE routing constraints.
We evaluate WaferTetris on Cerebras WSE-3 using LLM decoding and scientific stencil benchmarks. For LLM decoding, it achieves up to $4.42\times$ the system throughput of a template-based automatic mapping baseline on the same fabric size. Compared with the manually optimized WaferLLM, WaferTetris supports $1.8\times$ the context length for LLaMA3.1-8B (80k versus 44k), while delivering equal or higher per-user throughput at matched context lengths, up to $1.10\times$. - [498] arXiv:2611.18529 [pdf, html, other]
-
Title: eDIAR: Training-Free Open-Speaker Streaming Diarization for Long Meetings and Large Archives via e-meansComments: Demo & Project page: this https URLSubjects: Sound (cs.SD)
Meetings last up to an hour, and nobody knows in advance how many people will speak. Neural diarizers with a fixed number of output slots, such as Streaming Sortformer and Nemotron-3, are accurate and run online, but they cannot name more speakers than they have slots. We remove this limit without retraining. The slot head keeps deciding who is active and when speech overlaps, and speaker embeddings are clustered inside a single slot, with the official e-means algorithm, only after every slot has been used. In the streaming system, re-clustering is confined to a bounded window, labels are revised for a few seconds, and identities are anchored to the slots of the head. The resulting system has no limit on the number of speakers and stays accurate when speakers outnumber the slots, where fixed-slot models break down. On meetings that grow to an hour it stays far more accurate than Streaming Sortformer, which accumulates errors, and its cost per update does not grow with the meeting. The same recipe turns the four-slot Streaming Sortformer into an open-speaker system. Across recordings, e-means makes it affordable to link the speakers of a large archive frame by frame, a task that arises when conversational corpora are prepared, and it reaches the clustering quality of Lloyd's $k$-means several times faster. The method needs little tuning, and e-means runs with its default recipe from a few hundred to more than a million embeddings.
- [499] arXiv:2611.18530 [pdf, html, other]
-
Title: Understanding Contextual Bias in Agricultural CNNs Through Hidden-Neuron Concept InductionSubjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Convolutional Neural Networks (CNNs) can classify images using background-dominated patches that contain little or none of the target object, showing that they exploit contextual information. Previous studies detected this behavior through classification accuracy but say little about what CNN hidden layers represent. We investigate whether Concept Induction can reveal interpretable visual properties associated with hidden-neuron activations on contextual image patches. Using five agricultural datasets with high, moderate, and low contextual bias, we analyze four CNN architectures at three network depths and five patch locations. Quantitative image descriptors of each patch are encoded in a knowledge graph and used as background knowledge for Concept Induction, which derives logical expressions characterizing neuron activation. These expressions are tested on held-out images using the Mann-Whitney U test and Cliff's delta. Of 48,013 tested neuron-feature pairs, 13,164 (27.4%) were validated. Validation was most frequent at intermediate depths (mean 40.2% at mid layers versus 12.0% at deep layers), and no patch location dominated across datasets. Among the tested descriptors, edge, Gabor, and Tamura texture features had the highest validation rates. Stronger contextual bias did not correspond to higher neuron-feature validation rates: moderately biased Sugarcane had the highest rate (40.5%) and highly biased Cassava the lowest (19.2%). These results suggest that behavioral context dependence and the semantic interpretability of hidden neurons capture different aspects of CNN behavior, and that Concept Induction can complement patch-based bias detection.
- [500] arXiv:2611.18531 [pdf, html, other]
-
Title: Alignment by Socialization: Learning Tradeoffs Under SelectionSubjects: Artificial Intelligence (cs.AI); Populations and Evolution (q-bio.PE); General Finance (q-fin.GN)
Artificial intelligence (AI) agents are expected to complete tasks while following rules of conduct, but these objectives can conflict. We study how evolutionary selection shapes an agent's choices between them, and whether experience can make those choices reflect human judgments and withstand subsequent pressures. In our model, copies of an agent face a common uncertain reward state, and only those taking the rewarded action reproduce. We show that long-run growth is maximized when each action is taken as often as the environment rewards it. We compare three policies. A task-only policy thrives in a sandbox rewarding completion alone but fails when conduct matters. A fixed linear policy increasingly favors task completion as its stakes rise, even where humans never reward it. A learned policy instead converges to the tradeoffs rewarded by its environment; when humans supply the rewards, we call this socialization. Accumulated experience makes socialization harder to reverse under subsequent task-only training, although violations can arise before the prevailing disposition changes, and finite memory weakens this protection. When a single violation can end a deployment, a hard prohibition can prevent it, but a mistaken prohibition can also be costly. Experiments with small neural networks illustrate how training targets and memory shape these outcomes. Our results suggest that alignment depends not only on the objectives assigned to agents, but also on the environments that shape and sustain their behavior.
- [501] arXiv:2611.18533 [pdf, html, other]
-
Title: ModHarness-RL: Recasting Harnesses as Learnable and Evolving Programs for Agentic RLSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
Agentic reinforcement learning (RL) increasingly relies on a harness, the scaffold around the model that decides what it sees of its own interaction and what guidance it receives. Most existing approaches train under a hand-written harness that stays fixed throughout. However, an effective harness depends on the domain, the data and the model, and keeps changing as the model trains. We introduce ModHarness-RL, which recasts the harness as programs that are learned from data and evolve as the model trains with RL. ModHarness-RL records the interaction as structured State, a set of typed blocks, and two core components of the harness are programs over these blocks, a Compression program that assembles the accumulated history into the model's context and a Steering program that supplies guidance when behavior deviates. During RL training, candidate programs are proposed from recent trajectories, verified by execution, and replayed over past trajectories, so the harness tracks how the model's behavior shifts. We study CodeSearch in depth and further analyze ALFWorld and BIRD SQL, with models from 1.5B to 8B. The proposed Compression and Steering mechanisms improve held-out performance over the fixed-harness baseline in most settings. These results suggest that a harness for agentic RL can be learned from the task's data and evolved with the model.
- [502] arXiv:2611.18534 [pdf, html, other]
-
Title: Impact of Grid-Forming Inverter Placement and Penetration on Local and Aggregate FrequencySubjects: Systems and Control (eess.SY)
Modern power systems are undergoing a huge shift in generation mix, including inverter-based resources that are fundamentally changing the stability dynamics of the grid. Thus, synchronous Generators (SG) are gradually being displaced by Grid-forming inverters (GFM) that are proposed as a solution to continue supporting grid stability. However, the effect of GFM placement in the network on frequency performance and how well aggregate-frequency assessment captures placement sensitivity observed in local frequency metrics remains unclear. This paper evaluates all SG and droop-controlled GFM assignments in the IEEE 39-bus system. A reduced-order model is used for the exhaustive search with load disturbances applied separately at every eligible load bus while considering nadir and Rate of Change of Frequency (RoCoF) derived from both device rating weighted aggregate frequency and local bus frequencies. Results reveal that at fixed GFM penetration, placement produced up to 104 mHz variation in worst-case local nadir and 0.21 Hz/s in maximum local RoCoF. Aggregate-frequency assessment reduced the apparent across placements by 89\% for nadir and 88\% for RoCoF at 50\% GFM penetration. These results indicate that aggregate-frequency assessment alone may obscure favorable or adverse placement effects on local frequency performance.
- [503] arXiv:2611.18538 [pdf, html, other]
-
Title: Why Finite Intelligence Works in an Infinite World: A Mathematical TheoryComments: 56 pages, 5 figuresSubjects: Information Theory (cs.IT); Machine Learning (cs.LG); Machine Learning (stat.ML)
An information system often has to learn before it knows what it will be asked. A finite representation can succeed without recording an infinite environment in full when those asks share collective structure, rather than demand unrelated facts scattered arbitrarily across it. We formalize this phenomenon through future-task information, the minimum retained information needed to preserve decision value relative to the available evidence. A joint decision-regret state records posterior comparisons among actions for every permitted task. Under compactness assumptions, every retained representation admits a calibrated replacement based on this state, preserving all task regrets without increasing information; the resulting information-regret curve equals an operational coding rate. Two exact boundaries govern finite preservation: weak-star closure for a fixed action count, and the singular mass of the source-evidence law when all finite action menus are admitted. Collective structure determines information requirements beyond individual-question difficulty. Families with identical source laws, pairwise task metrics, and individual-task rates can have different high-resolution exponents. Covariance-identical coordinate and parity families have different exact rate-distortion curves, with optimal parity retention characterized by a sharp Fourier-entropy inequality. For balanced nested task families, refinement sizes and weighted reach determine the rate within an additive constant, attained by prefix coding under explicit decay conditions. Nonlinear learning and computational separations distinguish preservation from efficient discovery and use. The proposed mathematical theory identifies which shared decision requirements allow finite information to remain useful for infinitely many possible purposes.
- [504] arXiv:2611.18543 [pdf, html, other]
-
Title: Imitation Bootstrapped Multi-Agent Reinforcement LearningComments: 9 pages, 5 figures. Code, videos and results: this https URL. A preliminary version appeared at the PhysCollaBot Workshop at IROS 2026Subjects: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
Demonstrations are an effective way to accelerate single-agent reinforcement learning (RL), but extending this idea to multi-agent reinforcement learning (MARL) is non-trivial as it is difficult and often impossible for a single person to teleoperate multiple agents simultaneously. Round-Robin Behavior Cloning (R2BC) addresses this by letting a single demonstrator control one agent at a time while the others execute their current policies. However, the resulting imitation policies are limited by the quality of the demonstrations. We propose Imitation Bootstrapped Multi-Agent Reinforcement Learning (IBMARL), a framework that extends Imitation Bootstrapped Reinforcement Learning (IBRL) to cooperative teams and uses R2BC to train the imitation policies. At each time step, every agent receives one action proposal from its imitation policy and one from its RL policy, and an ensemble of centralized critics selects among the per-agent combinations of the two. A gated imitation penalty, a soft Q-filtered behavior cloning loss, regularizes each RL policy toward its imitation policy only where the critics value the imitation action more, so that after training each agent executes its RL policy alone, without the centralized critics. On four cooperative multi-agent RL benchmarks, IBMARL surpasses its imitation policies $2.6$-$4.2\times$ faster than demonstration-based baselines on two tasks, is the only method to surpass it on the other two, and exceeds the demonstrator itself on two of the four. With real human demonstrations IBMARL still surpasses imitation learning alone, and does so more than twice as fast as the best baseline.
- [505] arXiv:2611.18544 [pdf, html, other]
-
Title: Tight Bounds for k-Median++ and Improved Bounds for k-Means++Comments: 46 pages, 2 figuresSubjects: Data Structures and Algorithms (cs.DS)
The popular $k$-means++ algorithm, or $D^\ell$-sampling, is a fundamental primitive in both the theory and practice of clustering. In the bi-criteria setting, the algorithm samples $k+\Delta$ centers ($\Delta\ge0$), and its cost is compared with the optimum using $k$ centers. Despite extensive study over the past two decades, its exact approximation ratio and tight bi-criteria guarantees remain unknown.
We give a unified analysis for every distance exponent $\ell>0$, every $k\ge1$, and every $\Delta\ge0$. For metric $k$-median, our bounds are tight for all $k$ and $\Delta$. Without extra centers, the exact approximation ratio is $2(1+H_{k-1})$ (also in Euclidean space), improving the previous upper bound by a factor of $2$ ($H_j$ denotes the $j$th harmonic number). This tight analysis reveals that even one extra center improves the approximation ratio by an additive $\Theta(\log\log k)$, whereas previous bounds show no improvement.
For every $\ell>0$, our analysis gives upper bounds with the same dependence on $k$ and $\Delta$ as our tight bound for metric $k$-median, with leading constants depending only on $\ell$ and whether the metric is Euclidean. Beyond the improved dependence on $\Delta$, our leading constants are smaller than those in previous analyses. In particular, for the widely used Euclidean $k$-means objective, our upper bound without extra centers is $(1+\varphi)(1+H_{k-1})$, where $\varphi\approx1.618$ is the golden ratio, improving upon the previous constant of $5$. With oversampling, a constant fraction of extra centers is known to provide a constant-factor approximation. For example, with $10\%$ extra centers, our guarantee is $5.7$, compared with the best previous guarantee of $25.9$. - [506] arXiv:2611.18545 [pdf, html, other]
-
Title: From smoothed analysis to backward error: The power of adaptive perturbationsSubjects: Data Structures and Algorithms (cs.DS); Numerical Analysis (math.NA)
Smoothed analysis has become a central tool in theoretical computer science, often allowing for significant improvements to worst-case running times if we settle for a solution to a small random perturbation of the problem instance. This is closely related to the classical concept of backward error in numerical analysis, which allows an algorithm to return a solution to an arbitrary perturbation of the problem.
Despite the qualitative similarity of these concepts, we show that they can exhibit large quantitative differences. Notably, we prove that, while a linear system with any $n \times n$ matrix $A$ can be solved to backward error $\epsilon$ using $O(\log(n) / \epsilon)$ matrix-vector products with $A$, any algorithm that solves a random $\epsilon$-sized perturbation of $A$ requires $\Omega(n)$ matrix-vector products. This means that a useful perturbation for solving a linear system in backward error must be adaptive to the input. Interestingly, however, randomness is still essential: we show that any deterministic method that attains small backward error for all linear systems provably requires $\Omega(n)$ matrix-vector products, resulting in the first separation between randomized and deterministic algorithms in a natural computational model for solving linear systems. Finally, we show that matrix structure can significantly affect backward error complexity. For example, we show a deterministic $O(1/\epsilon)$ matrix-vector product algorithm for symmetric linear systems. - [507] arXiv:2611.18546 [pdf, html, other]
-
Title: Modeling the Frictional Contact of Continuum Robots with Environments through Linear Complementarity ProblemComments: Accepted for publication in International Journal of Robotics Research (IJRR). 21 pages, 18 figuresSubjects: Robotics (cs.RO)
Accurate friction modeling in continuum robots is essential for precise motion control and path planning. However, current approaches make predictions only for the mechanism's internal friction that is limited by the boundary forces of the Coulomb friction cone, which fail to characterize the static friction and motion hysteresis. In this paper, we propose a novel friction model based on complementarity constraints under the Coulomb friction assumption. This model predicts various friction effects caused by environmental interactions, including both static and sliding frictions and hysteresis behavior, as well as the corresponding robot deformation induced by these interactions, which are crucial for the practical motion planning and control of continuum robots. Additionally, real-time implementation of this model can be achieved, meeting the demand for computational efficiency in robotic applications. Experimental validations demonstrated that this model is able to reduce the maximum tip position error from 45.4 mm in the frictionless model to 5.85 mm in our friction-included model with a 287.9 mm-long tendon-driven continuum robot.
- [508] arXiv:2611.18549 [pdf, html, other]
-
Title: JVT: Just Video Transformer makes pixel-space video diffusion simpleComments: 23 pagesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Direct diffusion modeling of high-dimensional raw video data is challenging, and existing pixel-space methods often rely on specialized architectures. We introduce JVT (Just Video Transformer), a simple large-patch DiT-style Transformer for pixel-space video diffusion, without a pretrained autoencoder, auxiliary detailers, cascaded stages, or U-Net-style skip connections. We systematically examine video patch embeddings, register tokens, and adaptive layer normalization to identify design choices that balance generation quality and computational efficiency. The resulting JVT design achieves generation quality competitive with latent diffusion baselines in both class-conditioned and unconditional settings. The pixel-space formulation of JVT also naturally supports joint generation of pixel-aligned modalities. We demonstrate this capability through camera-conditioned RGB-disparity generation, which improves camera following relative to RGB-only baselines and yields explicit scene geometry without an additional reconstruction model at inference. These results highlight the potential of simple DiT-style Transformers for pixel-space video diffusion, and we hope our findings offer useful insights and practical guidance for future research.
- [509] arXiv:2611.18550 [pdf, html, other]
-
Title: Restoring Diversity in Distilled Image and Video Generators by Probing ActivationsFurkan Horoz, Ozgur Kara, Kaan Oktay, Adil Kaan Akan, Atakan Kara, Anand Bhattad, James Matthew RehgSubjects: Computer Vision and Pattern Recognition (cs.CV)
Distilling diffusion models into few-step student models substantially reduces sampling latency, but induces severe diversity collapse: samples generated from different random seeds show minimal variation. In this work, we investigate the root cause of this failure by probing the internal representations of student models relative to their teachers. We find that distilled students maintain total activation variance comparable to their teachers, but concentrate it into substantially fewer directions. This directional collapse is localized to an identifiable intermediate block at the initial denoising step. Consequently, diversity collapse stems from a misallocation of activation variance rather than an inherent loss of generative capacity. Guided by this finding, we propose ReDiverse to reinstate the missing variance directions. At training time, ReDiverse-FT fine-tunes a lightweight adapter that regresses student activations onto the teacher's, reviving the teacher's feature variance characteristic. At inference time, ReDiverse-TF introduces a training-free mechanism that steers activations along precomputed teacher principal directions. Across image and video distillation benchmarks, ReDiverse consistently restores sample diversity while preserving generation quality, further supported by the user study.
- [510] arXiv:2611.18553 [pdf, html, other]
-
Title: WeightBench: Evaluating Language Models as Weight-Space ProgrammersComments: 9 pages, 1 figure, 1 table. Code and leaderboard available at this https URLSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Recent work shows that large language models (LLMs) can operate directly on the parameters of other neural networks. They can read what a network was trained to do, fix a network that fails, and write working networks from scratch, all without gradient descent. This capability has been termed Weight-Space Programming (WSP). Exploring this new paradigm systematically requires a robust benchmark. We introduce WEIGHTBENCH, a benchmark that evaluates LLMs of increasing capability on 42 WSP tasks across five tracks: reading, fixing and writing weights, recognizing edits that are impossible, and data-free compression. The networks whose weights the LLMs read and edit range from 2K-parameter control policies to 7B-parameter language models. Experimental benchmark results show that frontier agents can already read small networks and write working control policies without ever running them, but cannot yet patch language models, compress them without data, or tell an impossible edit from a hard one. On all 42 tasks, Claude Opus 5.5 reaches a macro score of 0.535 (Sonnet 5.5 0.466, Haiku 5.5 0.402, DeepSeek v4-flash 0.215); on the 38 tasks it was run on, GPT-6.1-sol reaches 0.568 against Opus's 0.546. Code, the task registry, our OS-level sandbox harness, and the live leaderboard are publicly available at this https URL.
- [511] arXiv:2611.18559 [pdf, html, other]
-
Title: BAL: Bidirectional Autoregression in Latent Space for Learning Human Movement RepresentationsComments: Accepted to NeurIPS 2026. Project page: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Learning semantically meaningful human movement representations from 3D pose sequences is essential for human behavior modeling. Existing self-supervised methods derive supervision from artificial assumptions, e.g., invariance to augmentation within a certain range or recoverability of masked regions at a certain temporal scale, introducing heuristic hyperparameters that make the learned semantics sensitive to their settings. We propose BAL, a self-supervised framework that instead exploits supervision intrinsic to human movement: the natural ordering structure of everyday behavior. BAL autoregressively predicts in a semantically abstract latent space, analogous to how humans anticipate others' actions rather than their precise joint configurations. Bidirectional autoregressive prediction, both forward and backward in time, fully leverages these ordering constraints while enabling representation extraction using both past and future context at inference. To suppress representation collapse, BAL decomposes each prediction target into a global component capturing sequence-level semantics and a local component capturing context-independent movement. We prove that the local targets retain nontrivial angular diversity unless the representations are completely collapsed, preserving a meaningful training signal. Experiments on large-scale datasets demonstrate that BAL's representations are highly effective for motion captioning, forecasting, and interpolation, outperforming baselines on most evaluation metrics.
- [512] arXiv:2611.18562 [pdf, html, other]
-
Title: Training Reasoning Models in Simulation to Natively Use Real RobotsJiaming Tang, Zhuoyang Zhang, Junxian Guo, Fangzhou Zhao, Yilong Zhao, Zekai Li, Xingyang Li, Yuyang Zhao, Song Bian, Zhijian Liu, Song HanSubjects: Robotics (cs.RO); Machine Learning (cs.LG)
Robot-use agents (RUAs) have emerged as a promising approach to general-purpose robot control. However, strong performance has so far been achieved only by frontier proprietary models such as GPT-6 Astra and Claude Fable 5.2, whereas open-weight models remain substantially less capable. We show that an open-weight 27B reasoning model can be trained to use real robots natively through reinforcement learning, entirely in simulation and without demonstrations. Like a computer-use agent (CUA), the model observes the environment, reasons in a chain of thought, and generates the next end-effector pose, without any harness. We build a Post-training Engine for Agentic Robotics (PEAR) and use it to train the model on simulated robot tasks at scale with robot-use reinforcement learning (RURL), mirroring how frontier models acquire computer use. We find that the skills acquired in simulation generalize well beyond the training tasks: they transfer to an unseen robot, to the real world, and to novel instructions. After training on LIBERO and RoboDojo in simulation, the model succeeds in 60% of our trials on a real dual-arm robot, whereas the base model succeeds in none. On LIBERO, training raises the success rate from 1.6% to 74.4%, approaching GPT-6 Astra and GPT-6.1 Sol at maximum reasoning effort. Compared with $\pi_{0.5}$, a vision-language-action model fine-tuned on demonstrations, our model is substantially more robust to changes in the task itself. On LIBERO-Pro, $\pi_{0.5}$ succeeds in only 0.75% of episodes when assigned a new task in a familiar scene whereas our model succeeds in 57.4%. Finally, the model retains its general capabilities, with no measurable degradation on nine reasoning, coding and vision benchmarks. These results highlight robot-use reinforcement learning as a promising component of foundation-model post-training.
- [513] arXiv:2611.18563 [pdf, html, other]
-
Title: Pairing AIS with RF: Analysis of a Markovian Approach for Maritime Data FusionDaniel Grimm, Kyra Jongman, Renato Cordeiro Ferreira, Giuseppe Cascavilla, Alfredo Cuzzocrea, Willem-Jan Van Den HeuvelComments: 10 pages, 5 figures, accepted at the 13th IEEE Special Session Privacy and Security of Big Data of IEEE BigData 2026Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
Monitoring maritime activities is important for safety and security, particularly for detecting illegal activities, but remains challenging due to the size and complexity of the maritime domain. The Automatic Identification System (AIS) is commonly used for this purpose, despite its known vulnerabilities. In contrast, space-based radio frequency (RF) observations provide an independent source of vessel activity but lack ship identification and therefore cannot identify vessels on their own. Following the CRISP-DM process, this study applies a Pair Hidden Markov Model (PHMM) to align individual RF detections with AIS trajectories, using temporal and spatial information while modeling noise and timing discrepancies. We evaluate it on NOAA AIS data with synthetic RF detections against nearest-neighbor baselines and analyze its errors with bias tests on trajectory length, sampling density, stationarity and AIS--RF distance. The length-corrected Forward score reaches a top-1 precision of $0.7776$ on the validation split of multi-candidate detections, where errors are possible, and an AUC of $0.637$ for detecting dark vessels among the detections that pass the prefilter. With a single RF detection, the score mostly reflects the AIS trajectory itself, and its errors favor longer and more stationary trajectories. A log-odds score, which compares the full model with a null model without match transitions, raises precision to $0.8580$ and AUC to $0.852$. Its length-corrected variant no longer shows a significant bias towards stationary trajectories and short inter-message distances, but still favors candidates farther from the RF detection. A time-weighted nearest-neighbor baseline without fitting remains ahead (precision $0.9013$, AUC $0.904$), so this single-detection, hand-designed PHMM formulation approaches but does not beat it.
New submissions (showing 513 of 513 entries)
- [514] arXiv:2610.40863 (cross-list from math.CO) [pdf, html, other]
-
Title: The Cranston--Li--Su--Wang--Xu (2026) conjecture is trueComments: 15 pages, 5 figures, 1 table. Lean 4 proof: this https URLSubjects: Combinatorics (math.CO); Discrete Mathematics (cs.DM)
Let $A$ be a finite abelian group. A nowhere-zero $A$-flow of a graph can be changed along a cycle: going around the cycle, add a fixed non-zero element of $A$ to the value of every edge that points forward and subtract it from the value of every edge that points backward. The Cranston--Li--Su--Wang--Xu (2026) conjecture on the reconfiguration of nowhere-zero flows states that if $A$ has at least four elements, then in every graph in which all degrees are even, any two nowhere-zero $A$-flows are joined by a chain of such changes through nowhere-zero $A$-flows. It also states that, for every integer $k\ge 4$, the same holds for integer flows whose values are between $1$ and $k-1$ in absolute value, where each change adds a non-zero integer. We prove this conjecture. The proof is by induction on the number of edges. If a flow enters a vertex along one edge and leaves it along another edge with the same value, the two edges can be replaced by a single edge; this gives a smaller graph. The main step is to change every flow into a flow with such a pair of edges. For $A$-flows we do this when all degrees are at most $4$; repeatedly replacing a vertex of degree at least $6$ by a triangle, following Cranston et al., reduces the general case to this one. For integer flows, we show that, in a graph in which all degrees are even, any two nowhere-zero flows with values in the cyclic group $\mathbb{Z}_k$, $k\ge 4$, are joined by changes that add only $1$, $-1$, $2$ or $-2$, and that such changes lift to integer flows. All theorems of this paper are also proved in the Lean 4 proof assistant.
- [515] arXiv:2611.17343 (cross-list from physics.ins-det) [pdf, html, other]
-
Title: Curvature-Aware Edge Dynamic GNNs for Low-$p_T$ Track Reconstruction at HL-LHCSubjects: Instrumentation and Detectors (physics.ins-det); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); Data Analysis, Statistics and Probability (physics.data-an)
Particle track reconstruction in high-energy collider experiments is a computationally demanding combinatorial problem, traditionally solved by Kalman filters that scale poorly with increasing detector occupancy. Graph Neural Networks (GNNs) have recently emerged as a promising alternative. In this work, we introduce the Manifold Helical-IN, a novel Interaction Network architecture that explicitly embeds detector spacepoints onto a helical manifold chart using curvature-aware features. This geometric projection mathematically unrolls the tightly curling trajectories of low-momentum particles. Consequently, the network can distinguish true physical tracks from combinatorial noise with good precision. Evaluated on the TrackML challenge dataset, the proposed Manifold Helical-IN achieves a reconstruction efficiency of 0.9675 in the experimentally critical $p_T > 0.5$ GeV regime, approaching recently reported Transformer-based approaches. Crucially, the manifold embedding yields a fake track rate of 0.0037. This represents a 13-fold improvement over the baseline Helical-IN (0.0508). Furthermore, the model maintains good track coherence, as quantified by our proposed Mean Hit Recovery Fraction (MHRF) of 0.858, and achieves a highly efficient inference latency of approximately 150 ms per event on a single NVIDIA A100 GPU. These results represent a critical step toward real-time soft particle tracking for the High-Level Trigger (HLT). By operating within practical HLT latency budgets, this geometry-aware approach provides a pathway to recover soft, anomalous Beyond Standard Model signatures such as unclustered energy patterns (SUEPs), that routinely evade standard hardware filters.
- [516] arXiv:2611.17351 (cross-list from physics.ins-det) [pdf, html, other]
-
Title: Unsupervised Anomaly Detection in Measured Neutron Flux Profiles of the SAFARI-1 Research ReactorRethabile Kgolobe, Pavel M. Bokov, Simon H. Connell, Charis Harley, Bongani G. Maqabuka, Lesego E. MolokoComments: 33 pages, 26 figures, submitted to Annals of Nuclear EnergySubjects: Instrumentation and Detectors (physics.ins-det); Machine Learning (cs.LG)
Knowledge of the neutron flux distribution is essential for safe and effective reactor utilisation. In the SAFARI-1 research reactor, a set of axial flux profiles are periodically measured by activating natural copper wires inserted along the height of the fuel or control assemblies and subsequently measuring their activity using a sodium iodide detector. Improper wire insertion can, however, result in profiles that do not accurately represent the actual neutron flux distribution. Building on previously identified potential indicators of improper wire insertion, an automated approach based on unsupervised learning was developed to detect anomalous axial flux measurements. The measured data were preprocessed using missing-value imputation, decay correction, and normalisation. High-order polynomial fitting was then applied to characterise the axial flux profiles and derive features describing their shape. The Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm was employed to group flux profiles into regular and abnormal clusters based on these features with results matching the known groups. In addition, Convolutional Autoencoders were developed and trained on flux profiles from central and peripheral core positions to identify deviations from the regular profiles through reconstruction error, achieving better performance in central assemblies than in peripheral assemblies. The developed framework is intended to provide feedback to reactor operation by identifying potential wire-insertion problems in the manual technique used for axial flux monitoring at SAFARI-1.
- [517] arXiv:2611.17354 (cross-list from physics.ins-det) [pdf, html, other]
-
Title: Hybrid Quantum-Classical Sequence Models for Sample-Efficient Energy Regression in Simulated Calorimeter ShowersTariq Mahmood, Muhammad Ahsan, Talab Hussain, Grecia Itzel Colín-González, Alfredo Raya, Muhammad Awais Rafique, HassabAlla M.A. MahmoudSubjects: Instrumentation and Detectors (physics.ins-det); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); High Energy Physics - Phenomenology (hep-ph)
Energy reconstruction from electromagnetic calorimeter data underpins particle identification, event reconstruction and precision measurements in high-energy physics. Machine learning already performs well on this task, yet two questions remain open: how much the choice of data representation matters, and whether hybrid quantum-classical models add value in practice. We compare classical and quantum-enhanced models for incident positron energy reconstruction on the CaloGAN dataset, a public benchmark of 100,000 simulated electromagnetic shower events spanning 1 to 100 GeV. Eight architectures, namely SVR, ANN, CNN and LSTM together with their quantum counterparts, are trained on three calorimeter representations: engineered physical features, calorimeter images, and longitudinal shower sequences. Models that follow the longitudinal development of the shower reach the highest accuracy, which points to the value of physically informed representations. The hybrid QLSTM gives the lowest reconstruction error, with an RMSE of 0.294 GeV. Under matched-data conditions with only 1,500 training events, the QLSTM improves on the classical LSTM by 52.7% and also surpasses an LSTM trained on the full set. These observations indicate that longitudinal representations are central to calorimeter energy reconstruction and give preliminary evidence that hybrid quantum-classical sequence models can be more sample-efficient for regression in high-energy physics.
- [518] arXiv:2611.17356 (cross-list from physics.ins-det) [pdf, html, other]
-
Title: Waveform-based pulse timing enhanced by self-supervised neural regression models for digital nuclear instrumentationComments: 15 pages, 6 figuresSubjects: Instrumentation and Detectors (physics.ins-det); Machine Learning (cs.LG); Data Analysis, Statistics and Probability (physics.data-an)
With the rapid development of high-sampling-rate digital nuclear devices, their applications in nuclear facilities and related fields become more acceptable, and will possibly become widespread due to the versatility and flexibility of waveform sampling systems. When digitizers reach hundreds of megahertz or gigahertz, their lower limit in time measurement probes into the sub-nanosecond region, expediting their use in a bunch of time-critical scenarios. While conventional algorithms coupled with digital sampling techniques realize fair pulse timing in certain conditions, the potentials are still open to compact and efficient machine learning models. In this paper, we intend to boost the performance of pulse timing with a new paradigm of self-supervised learning on matched waveform pairs. The proposed method combines the mainstream self-supervised loss and the auxiliary weakly-supervised loss on the reference set, jointly improving the timing performance. We apply the method in a gamma/neutron-mixed radiation field, common in nuclear facilities. The simulation and experimental results show that, the proposed method not only outperforms conventional timing algorithms, but also surpasses neural network models based on supervised learning alone. Considering the trained model is lightweight and computationally efficient, it is promising to deploy it on edge devices, like economical microcontrollers with neural processing units.
- [519] arXiv:2611.17361 (cross-list from physics.ins-det) [pdf, html, other]
-
Title: Design and Experimental Evaluation of a Low-Cost Immersion Cooling SystemPramukto Mandal, Bhavya Jain, M. U. Kusheetha, Nikhil Meena, Ravindra Kumar, M. P. Sabarish, Amit Verma, Gaurav BhutaniComments: 20 pages, 12 figures, 3 tables. CAD files: this https URL ; firmware and data: this https URLSubjects: Instrumentation and Detectors (physics.ins-det); Systems and Control (eess.SY); Applied Physics (physics.app-ph)
This paper presents a low-cost, open-source, single-phase immersion cooling testbed for studying liquid cooling of data center hardware. A 500 W titanium heater, standing in for server load, is submerged in a glass tank of mineral oil, a dielectric fluid. The heated oil is circulated by a magnetic drive pump through an air-cooled radiator. An ESP32 microcontroller logs three DS18B20 temperature sensors (heater surface, bulk fluid and ambient) and switches the pump, fan and heater through a relay module. The complete system costs INR 35,318 (about USD 380), built entirely from components available in India. At quasi-steady state under the full 500 W load, the heater surface reached 118.3 °C while the bulk oil stayed at 29.3 °C, only 8.0 °C above ambient. Thermal-resistance analysis shows that the heater-to-oil interface, limited by natural convection, accounts for about 92% of the total resistance (178.0 vs 16.0 m°C/W). The external loop therefore has spare capacity. CAD files, firmware and data are openly available, so the setup can be reproduced and used to compare dielectric fluids, loads and flow arrangements.
- [520] arXiv:2611.17390 (cross-list from stat.ML) [pdf, html, other]
-
Title: Enumerating Posterior-Supported Priors: First-Order Error of Finite-Integral Cross-Categorical Measurement MorphismsSubjects: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
In generative modeling, standard measure theory adopts the a.e. convention, which legitimately ignores any error component carried by a null-support set: a single null element can hide out-of-domain error inside null support and never be explicitly accounted for. The contribution of this paper is not a new distributional distance, but rather a verifiable structure that writes down which underlying predicate supports a high-level object, when it is invertible, and at which layer the error resides, from which we propose the enumerating posterior-supported prior framework. The framework operates on a three-layer categorical chain Quot->MorTop->Set: the first two layers posit no point elements and take binary-predicate quotient partitions as objects; the replica-index counting measure successively discriminates one-to-many collections into injections, lifting object collection from surjection to a quotient-layer bijection; the mutual exclusion of preimages in the shared measure space is resolved by an invertible MPT. Four results consolidate into Thm. 7: binary quotient-partition meta-categorical objects support the set categorical functional along categorical support via measure descent--recovery, enumerating the morphism itself rather than its preimage and object; their multiplicity is normalized through replica-index injection and accounted as an explicitly first-order error. The contribution of this paper is a transformation from nontrivial-input construction -> engineering-computable first-order output: the binary quotient-partition meta-categorical objects, supported via categorical support measure descent--recovery, constitute a substantive nontrivial input---precisely because they lift the out-of-domain error that hid inside null support into a first-order falsifiable coefficient mu_err, a finitely obtainable finite integral first-order error.
- [521] arXiv:2611.17391 (cross-list from q-bio.BM) [pdf, html, other]
-
Title: Confident and Wrong: Structure Confidence Is a Poor Selection Filter for Protein FunctionComments: 12 pages, 5 figures, 4 tablesSubjects: Biomolecules (q-bio.BM); Artificial Intelligence (cs.AI)
Structure-guided protein design can generate plausible candidates, but structural confidence does not predict function. We trained a function-guided model on measured metal-binding affinity with protein sequence, ligand identity, and assay conditions, and used it to design rare earth element (REE) binding variants of lanmodulin, a lanthanide-binding protein with four EF-hand domains. We screened 297 designs from three strategies against terbium and dysprosium, adjacent lanthanides among the hardest REE pairs to separate. Selection on structural confidence inverted the outcome it was meant to improve. The Structure-Confidence-Selected panel carried the highest pTM, complex ipLDDT, and composite confidence scores, and expressed at 97%, yet only 16.7% of its designs bound terbium (95% CI 10.4 to 25.7), against 83.3% (74.6 to 89.5) for the function-guided panel, whose confidence scores were the lowest. The odds of binding differed 25-fold between the two panels (Fisher exact $p = 1 \times 10^{-20}$). No structural-confidence metric separated binders from non-binders consistently across panels, whereas the function-trained model's predicted pKd gave the highest ROC AUC in every panel, 0.79 to 0.92, and ranked measured affinity among function-guided binders at Spearman $\rho = 0.98$. On the design objective, Tb/Dy selectivity, the function-guided designs also exceeded the natural lanmodulin family, outranking a randomly drawn ortholog 84% of the time across 149 orthologs. Structural confidence is a plausibility signal, not a selection criterion for function; direct supervision on measured function is what ranks designs.
- [522] arXiv:2611.17449 (cross-list from physics.chem-ph) [pdf, html, other]
-
Title: Methanol substitution suppresses deflagration-excited knock and increases the efficiency of a hydrogen-fuelled CFR engineComments: 20 pages 8 figuresSubjects: Chemical Physics (physics.chem-ph); Systems and Control (eess.SY); Mathematical Physics (math-ph)
Knock restricts the performance of hydrogen spark-ignition engines and is usually attributed to end-gas autoignition. To the authors' knowledge, this is the first study in which methanol is used to replace part of the hydrogen in a hydrogen engine at constant fuel energy addition, and it shows that near the knock boundary the limiting oscillation is excited by the deflagration itself, which methanol removes without derating the performance. Port-injected methanol replaced 20--50% of the hydrogen energy in a variable-compression Cooperative Fuel Research engine at 600 rpm and compression ratios of 8--12. With pure hydrogen, oscillation in the 4--15 kHz range began before 10% of the charge had burned in 97--100% of cycles, and its maximum amplitude (MAPO) followed the maximum pressure-rise rate across 2499 cycles (Spearman $\rho=0.94$). Methanol slowed the deflagration: it replaced high-flame-speed hydrogen, cooled the intake by 24--25 K and, at unchanged intake pressure, enleaned the mixture by 10--13%. At 50% methanol, the 10--90% combustion duration lengthened 2.7--3.0-fold, oscillation onset moved 4.5--8.3$^\circ$CA after CA10 and mean MAPO fell by 89--93% although the spark timing was advanced by 13--19$^\circ$CA, and IMEP and indicated thermal efficiency rose by 2.4--7.8%. Methanol was thus worth more than this advance, and at matched CA50 MAPO was still about 90% lower. The measured combustion duration of all 25 data sets scaled with the computed laminar flame speed as $S_L^{-0.58}$ ($R^2=0.99$); fuel replacement accounted for about three quarters of the reduction in $S_L$, and the end-gas autoignition integral stayed below 0.07, against 1 at autoignition. The chamber kept its cylindrical acoustic modes, whose frequencies fell by 8--9%. The trade-off is formaldehyde emissions, which rose five- to sixfold while NO$_x$ fell by 49--85%.
- [523] arXiv:2611.17456 (cross-list from eess.SP) [pdf, html, other]
-
Title: Physical-Risk-Aware Decentralized Agents for Constrained Vehicle CommunicationsSubjects: Signal Processing (eess.SP); Multiagent Systems (cs.MA)
Decentralized V2X resource control is challenging because each vehicle has only local observations, while reliability depends on network-wide interactions. We develop a physical-risk-aware spatio-temporal graph attention (PA-STGAT) framework for agentic V2X control. Each vehicle perceives its local interaction graph, retains short-term temporal context, and independently selects its PC5 resource and power. Network outcomes are used during centralized training to refine the shared policy. Transmitter-specific credits map reliability failures to individual vehicles. A dual-priced graph term further converts network violations into coordination signals for discrete resource selection, with higher priority on short time-to-collision links. Experiments show higher raw-policy reliability than the tested graph-based baselines. Under filtered execution, PA-STGAT meets the nominal empirical communication requirements across all five seeds, with tradeoffs among utility, delivery reliability, resource occupancy, and filter-induced repair. Compared with GNN+MATD3, it requires fewer resource repairs but achieves lower utility and mean delivery reliability. Mechanism ablations show that removing dual-priced graph coordination causes resource-selection collapse. Density and mobility tests characterize the operating limits of empirical constraint satisfaction.
- [524] arXiv:2611.17468 (cross-list from eess.IV) [pdf, html, other]
-
Title: A thorough review on Deep Learning approaches to 2D-to-3D Registration in Medical ImagingAndrea Borghesi, Julia van Beurden, Maarten van Alphen, Michiel van den Brekel, Luc Karssemakers, Jan-Jakob Sonke, Baris Karakullukcu, Jonas Teuwen, George YiasemisComments: 51 pages, 4 tables, 8 figuresSubjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Medical Physics (physics.med-ph)
2D-to-3D image registration is a key component for several medical imaging purposes such as intraoperative guidance, enabling the alignment between preoperative 3D volumes and real-time 2D images. While conventional optimisation-based methods often struggle with non-convex objectives, sensitivity to initialisation, and high computational demands, deep learning (DL) has emerged as a promising alternative, offering improved alignment robustness and the potential for real-time performance. This review provides a comprehensive analysis of DL frameworks developed to address the challenges inherent in 2D-to-3D registration. We first outline how core concepts of DL are adapted to the registration problem. We then examine recent methodological advancements from multiple perspectives, including pipeline design, architectural innovations, and domain adaptation strategies. Particular attention is given to quantitative metrics, highlighting how performance is defined and measured across studies. Across the 41 included studies, we find that direct regression is the most common pipeline (23 studies, against 18 iterative ones), that hybrid CNN-attention architectures are gaining ground, and that research remains heavily concentrated on X-ray-to-CT registration. We conclude by identifying the absence of per-case uncertainty estimation and the strong neglect of temporal or sequential modelling as the two methodological gaps most consequential for clinical readiness, and discuss state-space models as a way to add temporal memory within intraoperative latency budgets. Further opportunities lie in more principled domain adaptation and in the integration of foundation models for representation learning.
- [525] arXiv:2611.17488 (cross-list from math.OC) [pdf, html, other]
-
Title: Convergence Properties of PF-Exec: Homogeneous Structure, Anchored Closed Form, and Gain DynamicsSubjects: Optimization and Control (math.OC); Machine Learning (cs.LG); Signal Processing (eess.SP)
This paper establishes a layered mathematical char acterisation of PF-Exec, the algorithm of the companion system paper. Demand-proportional bandwidth allocation yields an exact first-order homogeneous decomposition, from which the equal-load bandwidth target and its feasibility bound follow in closed form. The logarithmic knee view preserves strict load order and shares its global optimum with the Jain fairness index, while the demand-weighted centroid minimises the position potential. The damped softmax association block possesses a fixed point on the product of simplices. The truncated one-dimensional gain iteration admits a complete characterisation, with geomet ric convergence below the critical gain, initial-value-dependent bounded oscillation at it, and a family of bounded two-state cycles above it. A global invariant error ball and an explicit tracking recursion complete the execution guarantee. A numerical audit confirms the structural identities on every measured state and quantifies the local amplification of the coupled association and resource dynamics. Together, the results explain how closed form recovery directions and bounded adaptive gains provide stable execution without requiring a global Euclidean contraction assumption.
- [526] arXiv:2611.17489 (cross-list from math.CO) [pdf, html, other]
-
Title: Isolation of $k$-cliques in uniform hypergraphsComments: 13 pagesSubjects: Combinatorics (math.CO); Discrete Mathematics (cs.DM)
A copy of a hypergraph $F$ is called an $F$-copy. For any integers $k$ and $r$ with $2 \leq r \leq k$, let $K_k^r$ denote the complete $r$-uniform hypergraph whose vertex set is $[k] = \{1, 2, \dots, k\}$ (that is, the edges of $K_k^r$ are the $r$-element subsets of $[k]$). Given an $r$-uniform $m$-edge hypergraph $H$, the $K_k^r$-isolation number of $H$, denoted by $\iota(H, K_k^r)$, is the size of a smallest subset $D$ of the vertex set of $H$ such that the closed neighbourhood $N[D]$ of $D$ intersects the vertex sets of the $K_k^r$-copies contained by $H$ (equivalently, $H-N[D]$ contains no $K_k^r$-copy). We prove that if $H$ is connected and not a $K_k^r$-copy, then $$\iota(H, K^r_k) \le \left \lfloor \frac{(r-1)m + 1}{(r-1)(\binom{k}{r}+1) + 1} \right \rfloor.$$ The bound is best possible. This is a solution to a problem of Li, Zhang and Ye. We also provide an alternative proof for $r < k$. Both proofs introduce new ideas for establishing bounds on isolation parameters.
- [527] arXiv:2611.17491 (cross-list from math.SP) [pdf, html, other]
-
Title: A graph discretization of the Steklov and Neumann problems on domains in complete Riemannian manifoldsSubjects: Spectral Theory (math.SP); Numerical Analysis (math.NA)
We establish quantitative spectral convergence for graph discretizations of the Steklov and Neumann problems for bounded domains in complete Riemannian manifolds. Our construction uses weighted proximity graphs built from maximal $\epsilon$-separated sets, with a correction to the discrete energy accounting for the geometry near the boundary. We consider two discrete Steklov problems: one defined using an averaged boundary trace, and the other using restriction to boundary vertices. Both approximate the continuum Steklov spectrum, but under different conditions on the discretization parameters. The additional scale condition required for the restriction-based problem is also necessary for spectral convergence. As a byproduct, we obtain quantitative convergence of the Neumann spectrum. The estimates are uniform under suitable geometric bounds.
- [528] arXiv:2611.17503 (cross-list from stat.ML) [pdf, other]
-
Title: Multifidelity Operator Inference for Dynamical SystemsComments: 19 pages, AIAA SciTech Forum 2027 (Uncertainty Quantification Conference)Subjects: Machine Learning (stat.ML); Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Numerical Analysis (math.NA)
This paper presents a multifidelity (MF) formulation of Operator Inference (OpInf) for learning non-intrusive, physics-based reduced-order models (ROMs) of nonlinear dynamical systems. OpInf learns reduced operators from projections of snapshot data by solving a least squares regression problem. However, generating high-fidelity snapshots can be computationally expensive, limiting parameter coverage and the temporal extent of training trajectories for many aerospace applications, which can lead to inaccurate or unstable ROM predictions. This work proposes MF-OpInf, which combines limited high-fidelity snapshots with more plentiful low-fidelity snapshots through a control-variates-inspired MF least squares formulation to learn a single set of reduced operators across multiple trajectories. Low-fidelity trajectories provide additional parameter coverage and extend beyond the high-fidelity training time horizon. The low-fidelity snapshots are also used to construct an improved reduced basis through weighted proper orthogonal decomposition. We demonstrate MF-OpInf on a nonlinear reaction-diffusion problem with parametrized sinusoidal forcing and a two-dimensional incompressible flow past a cylinder with a parametrized inlet pulse. In these experiments, under an equal training simulation budget, MF-OpInf reduces the mean relative prediction error at unseen parameters by up to $87\%$ for the reaction-diffusion problem and $28\%$ for the cylinder flow problem over the test time regime, which extends beyond the high-fidelity training time horizon. These results demonstrate that MF-OpInf can improve temporal extrapolation and prediction at unseen parameters without increasing the training simulation budget.
- [529] arXiv:2611.17547 (cross-list from stat.ML) [pdf, html, other]
-
Title: Uncertainty-Aware Continual Learning for 2D XRD Interpretation Using Simulated and Experimental DataSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Autonomous experimentation systems require characterization models that can interpret heterogeneous measurements and maintain predictive performance as data accumulate. Developing such models involves addressing domain mismatch between data sources, overconfident predictions on unfamiliar inputs, and catastrophic forgetting during sequential updates. In this work, we present a proof-of-concept framework for uncertainty-aware continual interpretation of two-dimensional X-ray diffraction (XRD) data. The framework integrates established components: maximum mean discrepancy-based domain adaptation to reduce simulation-to-experiment discrepancies, a convolutional neural network with a spectral-normalized neural Gaussian process (CNN--SNGP) for predictive uncertainty estimation, and replay with knowledge distillation for sequential model updating. We evaluate the framework using simulated and previously-collected experimental data for crystal structure classification of single-phase body-centered cubic, face-centered cubic, and hexagonal close-packed patterns. Data are organized into predefined sequential blocks to assess adaptation and retention of previously learned knowledge. Within this setting, the results show accurate classification across simulated and experimental data, improved alignment between the two domains, and reduced forgetting compared with naïve fine-tuning. Predictive uncertainty is also evaluated on unfamiliar and noisy inputs. The contribution is the integration and evaluation of these capabilities for sequential XRD interpretation. The framework provides a basis for future integration into autonomous experimentation systems, with experimental planning, instrument control, and closed-loop deployment remaining areas for future development.
- [530] arXiv:2611.17561 (cross-list from physics.app-ph) [pdf, html, other]
-
Title: Contrastive local learning in large nonlinear analog circuitsComments: 19 pages (13 main, 6 supplementary) and 11 figures (5 main, 6 supplementary)Subjects: Applied Physics (physics.app-ph); Machine Learning (cs.LG)
Machine learning in physical hardware promises to reduce training cost by colocating computation, memory and adaptation within device physics. Contrastive local learning networks (CLLNs) offer a compelling route: an electrical circuit relaxes to equilibrium in free and weakly clamped states, and each adaptive element updates from local voltage contrasts rather than backpropagated error signals. Yet whether nonlinear transistor implementations can scale beyond proof-of-concept circuits, and what design principles control that scaling, have remained unknown. Here we show that nonlinear analog circuits trained by contrastive local rules can learn larger machine-learning tasks and remain robust under hardware nonidealities. Using device-level circuit simulations, we systematically explore initialization, operating regime, loss-dependent clamping and training parameters, and scale CLLNs beyond ten thousand adaptive edges. The same local rule trains classifiers that reach the accuracy regime of conventional digital methods on Ionosphere and handwritten-digit benchmarks, and a network that autoregressively generates language over a 122-token vocabulary. These results establish the feasibility of local learning in large nonlinear analog circuits.
- [531] arXiv:2611.17625 (cross-list from math.PR) [pdf, html, other]
-
Title: Finite speed Ornstein-Uhlenbeck-Kac processes and their applications in generative modelingSubjects: Probability (math.PR); Machine Learning (cs.LG); Analysis of PDEs (math.AP)
We study the Ornstein-Uhlenbeck-Kac (OU-K) process, where instead of a Brownian motion, a finite-speed Kac process defines the driving process. The OU-K process can be analyzed through a rescaled, time-inhomogeneous Kac process, characterized by a switching rate from the Kac part, and a convergence rate from the Ornstein-Uhlenbeck (OU) part. The law of this inhomogeneous Kac process describes the solution of a wave equation, the Euler-Poisson-Darboux (EPD) equation. For Kac switching rates being multiples of the OU convergence rate, we can explicitly derive this EPD solution via the sum of (the derivative of) two Legendre polynomials, restricted between two Dirac points representing the wave fronts. Using the framework of piecewise-deterministic Markov processes (PDMP), we describe the conditional velocity fields of the OU-K process via our polynomial-law-formula. This allows us to apply the OU-K process in a generative flow model. Since the OU-K velocity fields prove to be bounded by the finite Kac speed, the usual time truncations of diffusion models become unnecessary. We demonstrate advantages over diffusion and flow matching models in terms of generation sharpness on singular data sets, while achieving competitive results on CIFAR-10 image generation.
- [532] arXiv:2611.17634 (cross-list from cond-mat.mes-hall) [pdf, other]
-
Title: A Unified Device-Level Physical Framework for DRAM Read DisturbanceZhaoqiang Bai, Haocong Luo, Xinhe Wang, Xujing Li, Yuelin Jiang, Xuelu Huang, Guilei Wang, Hao Meng, Ataberk Olgun, Ismail Emir Yuksel, F. Nisa Bostancı, Jikun Wang, Onur MutluSubjects: Mesoscale and Nanoscale Physics (cond-mat.mes-hall); Hardware Architecture (cs.AR); Applied Physics (physics.app-ph)
Read disturbance in dynamic random-access memory (DRAM) depends on both local device interactions and the electrical conditions imposed by shared array circuitry. Here, we combine DIMM-level characterization results, device measurements, and TCAD simulations to interpret RowHammer, RowPress, and ColumnDisturb within a unified device-level physical framework. We relate charge evolution at the floating storage node to three key mechanisms: trap-assisted electron migration (EM), capacitive coupling (CC), and sense-amplifier-mediated disturb (SAMD). EM is local and trap-sensitive, while CC weakens with wordline separation. SAMD extends disturbance through shared column circuitry by modifying unselected-cell leakage under rail-level bitline biases, a contribution supported by the approximately ninefold higher OFF-state leakage measured relative to retention-like conditions for a high-voltage storage state. These distinct spatial ranges account for localized RowHammer and RowPress failures and remote ColumnDisturb populations with abrupt changes at subarray boundaries. We further interpret prolonged RowPress activation as a transition from an initial RowHammer-like transient to sustained leakage driven mainly by SAMD under held wordline and bitline biases. Within this device-level framework, trap effectiveness and storage-state-dependent leakage provide an interpretation of previously puzzling failure signatures, including the earlier onset of 0-to-1 RowHammer failures despite the predominance of 1-to-0 failures at high hammer counts. The same framework explains the observed RowPress and ColumnDisturb polarities and the duration-dependent reversal of neighboring- and passing-wordline effectiveness. These results provide a physical basis for developing compact DRAM reliability models across access patterns and array configurations.
- [533] arXiv:2611.17678 (cross-list from math.CO) [pdf, html, other]
-
Title: Nonexistence of a quasi-symmetric $2$-$(33,15,35)$ designSubjects: Combinatorics (math.CO); Information Theory (cs.IT)
It is shown that there exists no quasi-symmetric $2$-$(33,15,35)$ design with intersection numbers $6$ and $9$. This is proved by considering the ternary code generated by the incidence vectors of the blocks and its dual code.
- [534] arXiv:2611.17682 (cross-list from math.OC) [pdf, html, other]
-
Title: Agent-written heuristics versus commercial solvers on unsolved job-shop benchmark instancesComments: 27 pages, 4 figures, 14 tables. Submitted to International Transactions in Operational Research. Data, schedules, checkers and code: this https URLSubjects: Optimization and Control (math.OC); Artificial Intelligence (cs.AI)
Commercial solvers are the standard tool for difficult scheduling problems. We study whether a frontier large language model, used as a coding agent with open-source tools only, can match them in quality and speed. We report every run, and independent checkers verify every schedule. On the 100,000-operation job shop tai_100_1000_1, an agent-written solver reached 525,515, below OptalCP, Hexaly, and CP Optimizer. Unchanged, it improved all ten instances of the family and closed two. With equal budgets on the same workstation, it ended 0.0-1.8% above the lower bound, against 2.6-5.8% for CP Optimizer and 2.4-4.4% for Hexaly. On the Behnke-Geiger flexible job shops, an agent-written tabu search improved 11 of 32 open instances. With two transport robots, an agent-written solver improved seven of ten MKT instances. New lower bounds show that four published MKT values cannot be attained on the distributed instance files. Agent-run improvement campaigns closed eight further instances.
- [535] arXiv:2611.17712 (cross-list from eess.IV) [pdf, html, other]
-
Title: A Tumour-Aware Multimodal Framework for Explainable Prediction of IDH Mutation in Gliomas from Multi-Sequence MRIComments: 45 pages, 12 figures, 4 tables. Preprint of an article submitted for consideration in International Journal of Information Technology & Decision Making (World Scientific)Subjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
Accurate prediction of isocitrate dehydrogenase (IDH) mutation status is essential for glioma diagnosis, prognosis, and treatment planning. Although recent deep learning approaches have demonstrated promising performance, many rely on whole-brain feature extraction and provide limited interpretability. This study proposes a tumour-aware multimodal framework for explainable, non-invasive prediction of IDH mutation status from multiparametric magnetic resonance imaging (MRI). The proposed framework integrates four MRI sequences (T1, T1CE, T2, and FLAIR), tumour subregion segmentation, quantitative radiomic descriptors, and clinical information within a unified deep learning architecture. An ablation study was conducted to evaluate the contribution of tumour-aware learning, radiomic features, and clinical variables. Model performance was assessed using multiple complementary evaluation metrics, including accuracy, precision, recall, F1-score, balanced accuracy, receiver operating characteristic analysis, confusion matrix evaluation, and Grad-CAM visualisation. Experimental results demonstrated that the proposed multimodal framework consistently outperformed the baseline MRI-only model, with the Clinical Fusion model achieving the strongest overall predictive performance. The integration of tumour-aware learning, radiomic descriptors, and clinical information progressively improved classification performance, while Grad-CAM visualisations indicated that the model predominantly focused on biologically relevant tumour regions. These findings suggest that tumour-aware multimodal feature integration enables accurate, interpretable, and non-invasive molecular profiling of gliomas and provides a promising foundation for future clinical decision support in precision neuro-oncology.
- [536] arXiv:2611.17727 (cross-list from math.OC) [pdf, html, other]
-
Title: Proximal Adam algorithm for nonsmooth composite optimization and linear convergence under proximal PL geometryComments: 27 pages,3figuresSubjects: Optimization and Control (math.OC); Numerical Analysis (math.NA)
We analyze Adam-type methods for nonsmooth composite optimization problems $\min_x\Phi(x)=f(x)+h(x)$, where $f$ is smooth and $h$ is convex, proper, and lower semicontinuous. The nonsmooth term is treated by a metric-weighted proximal map, while the smooth term is preconditioned by the Adam second-moment matrix. We prove a composite descent lemma whose constants depend explicitly on the spectral bounds of the adaptive metric and a residual--step equivalence that converts proximal step decay into subgradient stationarity. These yield an $O(1/N)$ bound on the minimum proximal stationarity residual for the deterministic proximal normalized gradient method without convexity of $f$, and the same $O(1/N)$ rate for proximal Adam under quantitative momentum-alignment and residual-coherence conditions together with a metric-compatibility condition. By estimating the momentum error directly, the same $O(1/N)$ rate is also obtained without the momentum-alignment and residual-coherence assumptions. When the composite objective satisfies a proximal Polyak--Łojasiewicz inequality, both algorithms and the relaxed proximal Adam scheme converge linearly with explicit contraction factors. Numerical experiments on a diagonal Lasso problem, two-dimensional image TV denoising, and two-dimensional elastic parameter identification illustrate the predicted stationarity and linear rates.
- [537] arXiv:2611.17736 (cross-list from math.OC) [pdf, html, other]
-
Title: High-probability convergence of adaptive, proximal, and clipped Adam-type methods for nonconvex stochastic optimizationComments: 30 pages, 7 figuresSubjects: Optimization and Control (math.OC); Numerical Analysis (math.NA)
This paper develops a unified high-probability analysis of adaptive, proximal, and clipped first-order methods for nonconvex stochastic optimization. For a smooth objective whose stochastic gradient error has sub-Gaussian norm, the normalized adaptive gradient method attains the standard stochastic nonconvex complexity of the square root of the number of iterations, with an additional logarithmic dependence on the inverse confidence level that is inversely proportional to the number of iterations. The proof uses an adaptive descent decomposition together with Freedman and Bernstein concentration inequalities. The same guarantee is established for the full adaptive moment method under an explicit momentum-alignment condition. The analysis is extended to projected and proximal updates through a stationarity measure based on the subdifferential distance, to objectives satisfying the Polyak-Lojasiewicz condition, and to heavy-tailed noise through truncation. The framework is applied to stochastic inverse problems in elasticity, and numerical experiments illustrate the theoretical scaling and the practical behavior of the methods.
- [538] arXiv:2611.17743 (cross-list from eess.IV) [pdf, html, other]
-
Title: X-Strata: Separating Aneurysms and Overlapping Vessels into Layers from 2D X-ray AngiographyZhihong Wen, Congzhou Chen, Hucheng Li, Xun Han, Jianning Zhang, Yakun Chen, Hao Gao, Hao Zhao, Yudong Ma, Guanqiao Li, Yongping Liang, Shengli GuoComments: 22 pages, including 10 pages of supplementary materialSubjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
An X-ray angiogram sums every structure along a ray into one pixel. When vessels cross an aneurysm, they obscure part of it and mix their contrast into its filling curve. We present X-Strata, a network that separates a cine angiogram into three layers: the complete aneurysm sac, the vessels in front of it, and the vessels behind it. Contrast timing indicates how much each layer contributes to a pixel, and parallax between two views indicates their order, which a single view cannot determine. X-Strata first estimates the layers and their contrast curves, then uses the second view to assign vessels to the front or back layer, and lifts all layers into 3D. It learns from layer labels projected from same-session 3D rotational angiography, but needs only 2D cine runs and their acquisition geometry at inference. On 220 held-out patients from three centers (122 external), two-view X-Strata ranks first among 14 baselines, including VGGT and C$^2$RV, in recovering the sac where vessels overlap it (Dice 0.693 internal, 0.641 external) and in ordering the layers (balanced accuracy 0.703, 0.659). Its order accuracy exceeds that of all 11 end-to-end baselines by at least 4.1 points internally and 3.3 externally (Holm-adjusted $p < 0.001$) and grows with view separation. Recovered 3D sac surfaces have a median Chamfer distance of 0.34 mm on the internal cohort. X-Strata may therefore provide layered aneurysm and vessel information, including 3D sac shape, for patients who undergo only 2D angiography.
- [539] arXiv:2611.17745 (cross-list from math.OC) [pdf, html, other]
-
Title: The Polyak-Łojasiewicz Paper: A 10-Year RetrospectiveComments: 20 pages, 6 figuresSubjects: Optimization and Control (math.OC); Machine Learning (cs.LG); Machine Learning (stat.ML)
In 2016, we wrote a paper introducing the phrase "Polyak-Lojasiewicz (PL) condition" for an old idea, unsure of how readers would react. The paper ultimately had a larger impact than we expected, and we are surprised and honoured to receive the 2026 PKDD ECML Test of Time Award. With 10 years of hindsight, this note offers an informal and candid retrospective on the PL condition. We take a look back at the origins of the paper, give an AI-assisted refined version of the main theorem, review criticisms of the work, and discuss a variety of related ideas and extensions.
- [540] arXiv:2611.17804 (cross-list from quant-ph) [pdf, html, other]
-
Title: Tight Quantum Query Bounds for Collision Finding over Arbitrary Bounded TypesSubjects: Quantum Physics (quant-ph); Computational Complexity (cs.CC)
We determine the bounded-error quantum query complexity of finding one $r$-collision in an $N$-input function of prescribed type, which specifies all equality-class sizes. For every fixed maximum class size and collision order, we give a tight bound as the maximum of scales determined by the numbers of classes of size at least $2,\ldots,r$. The lower bound holds for a uniformly random function of that type, and a matching algorithm succeeds on every such function. No intermediate class size is required to occur, and every feasible output alphabet is allowed.
The characterization connects the abundant-collision regime associated with set equality to the sparse-target regime of $k$-distinctness. In particular, finding the unique $(k+1)$-collision in type $(k^n,k+1)$ has complexity $\Theta_k(N^{3/4-1/[4(2^{k+1}-1)]})$, where $N=kn+k+1$, resolving an open problem of Belovs.
For the lower bound, we introduce a filtered compressed permutation oracle that projects onto a harmonic database subspace after each complete function query. Its $O(T/N)$ comparison error over $T$ queries overcomes the square-root limitation of the unfiltered process. Hidden capacity extensions and moment estimates retain the populations at each intermediate collision order. Matching upper bounds use sampled collision lists and colored adaptive learning graphs, with a coherent decision-to-search conversion that avoids logarithmic overhead. We also give an alternative nested quantum-walk algorithm for three-collision search, matching the bound up to polylogarithmic factors. - [541] arXiv:2611.17854 (cross-list from physics.ao-ph) [pdf, other]
-
Title: Tracing Uncertainties of Extreme El Niño Prediction to Mesoscale Ocean CurrentsBo Qin, Zeyun Yang, Qiaofei Xue, Yanan Ding, Jinyu Wang, Chaopeng Ji, Shijin Yuan, Bin Mu, Wansuo Duan, Guokun Dai, Mu Mu, Yi DengComments: 30 pages, 13 figuresSubjects: Atmospheric and Oceanic Physics (physics.ao-ph); Artificial Intelligence (cs.AI)
The predictability of the El Niño-Southern Oscillation (ENSO) has historically been linked to large-scale errors in models' initial conditions such as heat content and surface wind stress. Here we demonstrate that these slowly evolving errors may ultimately originate from mesoscale errors in initial oceanic states that are common in current models and ocean observation systems. This hidden source of ENSO forecast uncertainty, in terms of mesoscale ocean-current errors, is convincingly exposed by the conditional nonlinear optimal perturbation (CNOP) method applied to a differentiable ocean AI model. The CNOP-identified most influential current errors are strongly event-dependent. For the 1997/98 El Niño, the initial current errors in the eastern Pacific results in the largest forecast error through locally suppressing original subsurface warming. For the 2015/16 event, the dominate error originates from the western Pacific, which reorganizes upper ocean temperature structure and affects the eastern Pacific through the equatorial waveguide. Further analysis reveals that these current errors perturb mesoscale eddy count and eddy kinetic energy triggering large-scale temperature errors that hinder El Niño development. The timing of the maximum growth rate of the obtained CNOP error is well aligned with the spring predictability barrier of ENSO, and the error structures are reflected in simulated analysis errors caused by unresolved mesoscale currents. These findings extend the potential source of ENSO predictability beyond large-scale thermal signals, reveal for the first time an uncharted route of ENSO forecast failure, and provide an important physical basis for targeted ocean observations and data assimilation in the future.
- [542] arXiv:2611.17884 (cross-list from math.LO) [pdf, html, other]
-
Title: HOL in HOL: mechanized soundness, completeness and consistency, with and without infinityChristoph Benzmüller (Otto-Friedrich-Universität Bamberg and Freie Universität Berlin), Daniel Kirchner (Otto-Friedrich-Universität Bamberg)Comments: 23 pages, 1 figure. Formal proof artifacts: Archive of Formal Proofs entries HOL_in_HOL_Deep and HOL_in_HOL_Deep_Infinity_ZFC (Isabelle2026). SubmittedSubjects: Logic (math.LO); Logic in Computer Science (cs.LO)
What can a logic certify about a copy of itself? For Church's simple theory of types, formalized as a deep embedding in Isabelle/HOL with Henkin semantics and a natural-deduction calculus, the answer has two tiers. On the first tier, plain Isabelle/HOL certifies the logic without an axiom of infinity, by a standard model with one individual. On the second tier, the logic with a Dedekind-style axiom of infinity is certified only in Isabelle/HOL extended by a large enough set-theoretic universe. Both tiers appear in Harrison's self-verification of HOL Light. What the present formalization adds is Henkin completeness, and with it a semantics that matches the calculus exactly, compactness and model existence on the first tier, and the boundary between the tiers. Comparing two ways of adding infinity to the object logic shows where the boundary lies. One is an infinite domain of individuals, expressed by the first-order sentences "there are at least $n$ individuals", each of which has a finite model; the other is the Dedekind axiom, a single sentence with no finite model, and there the first tier ends. In the calculus the axiom is strictly stronger than the sentences together: it derives each of them, but they do not derive it. The witness is a countable Henkin model in which every sentence holds and the axiom fails, because its function domain is too sparse for an injective, non-surjective self-map. On the second tier, the universe serves one purpose only: to prove the Dedekind axiom consistent. Everything that follows, completeness and even the countable Henkin model of the axiom, is obtained in plain Isabelle/HOL. The universe axiom is Harrison-style, realized in Paulson's ZFC_in_HOL with the domain of every type in $V_{\omega+\omega}$; with it, full classical higher-order logic with infinity and relational choice is consistent relative to set theory.
- [543] arXiv:2611.17971 (cross-list from stat.ML) [pdf, html, other]
-
Title: Efficient and Practical Clustering of Variables in Causal GraphsSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
Clustering variables in a graph may improve visual clarity and simplify causal inference. Harnessing the full potential of clustering requires an automated method to detect useful clusters in a graph. We introduce a novel class of clusters called conduits, and present a sound, complete and efficient algorithm for detecting all conduits in the graph. In addition to algorithmic detectability, conduits appear naturally in causal graphs stemming from real-world applications, enabling substantial size reduction. Furthermore, the structure of conduits enables the application of known identifiability results regarding clustering. We demonstrate these properties in a simulation study with real-world graphs.
- [544] arXiv:2611.17974 (cross-list from eess.IV) [pdf, other]
-
Title: Few-Shot Open-Set Hyperspectral Pixel-Wise Classification with Spatial Aware Label PropagationJournal-ref: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2026, 19, pp.30813-30836Subjects: Image and Video Processing (eess.IV); Neural and Evolutionary Computing (cs.NE)
Pixelwise classification of hyperspectral images (HSI) is a fundamental task in remote sensing, but its deployment in real-world applications is often limited by the scarcity of labeled data and the presence of unknown classes at inference time. In this work, we propose Few-shot Open-set hyperspectral recognition from Self-Supervised Learning training (FOSSeL), a unified framework for few-shot, open-set hyperspectral pixelwise classification that does not require supervised training. The proposed approach first leverages self-supervised learning to construct a discriminative feature space from unlabeled data. Then, given a small set of labeled seed pixels, class predictions are obtained through spatially-aware graph-based label propagation over superpixel representations, enabling the joint exploitation of spectral similarity and spatial consistency. To address the open-set scenario, we introduce simple yet effective strategies to identify pixels that do not belong to the known classes. We evaluate FOSSeL on three benchmark hyperspectral datasets under multiple few-shot settings. The results show that the proposed framework achieves competitive or superior performance compared to fully supervised methods while maintaining robustness in the presence of unknown classes. These findings highlight the potential of combining self-supervised feature learning with graph-based inference for practical HSI analysis in low-annotation and open-world scenarios.
- [545] arXiv:2611.17987 (cross-list from math.OC) [pdf, html, other]
-
Title: The Train Unit Shunting Problem with Service Scheduling and Workforce Assignment: classical and quantum optimization approachesComments: Train Unit Shunting Problem, Railway Optimization, Mixed-Integer Linear Programming, Quantum Annealing, Hybrid Quantum-Classical OptimizationSubjects: Optimization and Control (math.OC); Emerging Technologies (cs.ET); Quantum Physics (quant-ph)
Efficient railway rolling stock management requires the coordinated planning of train-unit assignment, shunting movements, track occupation, service operations, and workforce allocation. These interdependent decisions give rise to a challenging combinatorial optimization problem, particularly as the size of the fleet and the planning horizon increase. This paper introduces the Train Unit Shunting Problem with Service Scheduling and Workforce Assignment, a mixed-integer linear programming formulation that integrates these operational decisions within a unified optimization framework. To improve computational tractability, a preprocessing procedure is developed to reduce the size of the resulting model. The proposed approach is evaluated through computational experiments on instances of increasing size and under different objective-function weight and cost configurations. Classical optimization experiments using Gurobi show that filtering method reduces the number of decision variables and constraints, while its impact on computational time varies across instances. In addition, a constrained quadratic model representation is investigated using D-Wave's hybrid quantum-classical optimization solver.
- [546] arXiv:2611.17999 (cross-list from math.CO) [pdf, html, other]
-
Title: Additive cubes are avoidable over $\{0,1,2,3\}$, $\{0,1,4\}$ and $\{0,2,5\}$Subjects: Combinatorics (math.CO); Discrete Mathematics (cs.DM); Formal Languages and Automata Theory (cs.FL)
We prove that additive cubes are avoidable over the alphabet $\{0,1,2,3\}$, thereby solving a long-standing question initiated by Pirillo and Varricchio in 1994. By extending the reduction and spectral machinery of Cassaigne, Currie, Schaeffer, and Shallit (2014), we prove that the search for additive cubes in such morphic images reduces to a finite graph traversal. We exhibit eight explicit morphisms from $\{0,1,3,4\}^*$ to $\{0,1,2,3\}^*$ of length 3 whose images on the Cassaigne \textit{et al.} infinite word contain no additive cubes.
- [547] arXiv:2611.18012 (cross-list from physics.comp-ph) [pdf, html, other]
-
Title: ADfiRe: An empirical- and CFD-informed advection-diffusion-reaction thermal model for wildfire spreadKonstantinos Vogiatzoglou, Nikolaos Sdiros, Costas Papadimitriou, Konstantinos Ampountolas, Petros Koumoutsakos, Vasilis BontozoglouComments: 35 pages, 9 Figures, 2 TablesSubjects: Computational Physics (physics.comp-ph); Computational Engineering, Finance, and Science (cs.CE); Mathematical Physics (math-ph); Fluid Dynamics (physics.flu-dyn)
Real-time wildfire forecasting requires a trade-off between rapid yet physically relevant models and high fidelity but computationally costly models of flame spread over heterogeneous, multiphase solid fuel. The present study introduces ADfiRe, a hybrid framework that integrates empirical correlations and CFD-informed parameterizations into a physics-based advection-diffusion-reaction (ADR) formulation. The key features of ADfiRe are the use of established fuel databases, the capacity to handle multiple fuels, and a methodology to capture the effect of arbitrary topography. Furthermore, it includes the effect of flame tilting due to wind, thus accounting for the differences between plume- and wind-dominated fire propagation. ADfiRe resolves explicitly the coupled thermochemical degradation as well as heat and mass transport mechanisms of the advancing fire front. The wind field is modeled based on an appropriately defined Byram number. The model is validated against five well-documented experimental field-scale fires in Mediterranean shrublands, showing quantitative agreement with key metrics such as the rate of spread (ROS) and the fireline intensity, through a two-stage, Arrhenius-based heterogeneous reaction scheme coupling solid-fuel pyrolysis and oxygen-limited char/volatile combustion.
- [548] arXiv:2611.18066 (cross-list from nucl-th) [pdf, html, other]
-
Title: Product units in gated recurrent networks enhance prediction of nuclear masses and charge radiiSubjects: Nuclear Theory (nucl-th); Machine Learning (cs.LG)
The prediction of nuclear properties using machine learning can complement theoretical models and advance the exploration of poorly known domains of the nuclear chart. We propose a machine learning technique for predicting nuclear masses and charge radii based on gated recurrent units (GRU) that integrates learnable product-unit mappings into the candidate state to model multiplicative dependencies among neighboring nuclei. Computations are performed in the complex domain to jointly capture amplitude and phase dynamics. Across three data partitions and three training seeds per partition, AM-PU-GRU achieves a mean mass interpolation RMSE of 0.2698 MeV and a temporal RMSE of 0.1905 MeV. Configuration comparisons support the integration of learned products into gated recurrence. Without mass inputs, MI-PU-GRU predicts charge radii with an RMSE of 0.019947 fm, 14.40% below the fixed WS* prior. Matched radius controls further show how candidate fusion shapes the benefit of learned products. Paired errors, nuclide-resolved maps, and inference measurements characterize predictive gains and computational cost. This article is an extended version of our ICCS 2026 conference paper (doi:https://doi.org/10.1007/978-3-032-29921-5_10).
- [549] arXiv:2611.18095 (cross-list from stat.ML) [pdf, html, other]
-
Title: Training-point consistency of kernel ridge regression with varying correlation scalesSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST)
The scale parameter in kernel ridge regression plays a pivotal role in modeling the correlation structure of data. In modern applications, sampling density and the underlying correlation range often change simultaneously, making a fixed-scale assumption highly restrictive. However, establishing asymptotic theories under varying scale parameters is fundamentally challenging. To address this gap, this paper develops a unified theoretical framework for kernel ridge regression under varying scale parameters. We establish that, on uniform tensor grids with product exponential kernels, the consistency of the training-point risk holds if and only if the effective local sample size diverges to infinity. This sharp criterion universally covers all regimes along arbitrary deterministic scale sequences, seamlessly unifying the long-range, fixed-scale, and short-range. Furthermore, we provide an explicit, signal-independent construction of a deterministic regularization penalty that guarantees this consistency. Numerical experiments illustrate the criterion and the corresponding penalty across all the three regimes.
- [550] arXiv:2611.18105 (cross-list from q-bio.NC) [pdf, html, other]
-
Title: An In-Silico Audit of Topographic Organization in a Multimodal Brain-Response ModelComments: 12 pages, 5 figures, 4 tables. Code and stimuli: this https URLSubjects: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI)
TRIBE v2 (d'Ascoli et al., 2026) predicts fMRI responses across the cortical surface from video, audio and text. Its authors validated it with category-selective localizers (faces, places, bodies, words) and language contrasts, but not with the classic topographic organization of sensory cortex: retinotopy, tonotopy, and motion selectivity in area MT+. There is a clear architectural reason to expect retinotopy to fail: TRIBE's video features are averaged over space before the response head, which should discard visual-field position. I present mapping stimuli, a fitting pipeline adapted from standard fMRI mapping analyses, and a stand-in model with known ground truth that serves as a positive and negative control, and apply them to TRIBE v2's predictions, comparing the fitted maps with human atlases using spin tests. TRIBE v2 shows a sharp split. There is no contralateral hemifield response and no polar-angle map, robustly across stimulus controls. There is no tonotopic gradient in primary auditory cortex: best frequencies span an interquartile range of only 0.14 octaves across core and belt, and the null persists with loudness-equalized tones. An eccentricity-like gradient appears only when the ring geometry changes, consistent with global image statistics rather than a positional code. In contrast, motion selectivity is strong and precisely localized: 22% of the 1% most motion-selective vertices fall in MT+, which covers 0.42% of cortex. In-silico experiments that depend on feature identity appear viable with TRIBE v2, while those that depend on visual-field or frequency position are not supported by the current model.
- [551] arXiv:2611.18124 (cross-list from math.OC) [pdf, html, other]
-
Title: Muon Attains $\mathcal{O}(\varepsilon^{-3})$ Iteration Complexity under ShufflingComments: 25 pages, 14 figuresSubjects: Optimization and Control (math.OC); Machine Learning (cs.LG)
Muon is a stochastic first-order optimizer for matrix-valued parameters. It maintains an exponential moving average of stochastic gradients in a momentum buffer with coefficient $\beta$, and orthogonalizes the buffer to obtain the update direction. Muon has been reported to accelerate and stabilize the training of large language models. In practice, it is often used with shuffling (sampling without replacement), where each epoch visits the components of a finite-sum objective in a permuted order. However, its iteration complexity in this setting remains open. We analyze Muon under shuffling, with the momentum buffer orthogonalized by the polar map. For smooth nonconvex finite-sum problems, we establish an $\mathcal{O}(\varepsilon^{-3})$ iteration complexity for finding an $\varepsilon$-stationary point. This guarantee holds for every sequence of permutations, covering random reshuffling, shuffle once, and the incremental gradient method. This complexity matches that of stochastic gradient descent under shuffling in its dependence on $\varepsilon$, and improves on Muon's $\mathcal{O}(\varepsilon^{-4})$ complexity under sampling with replacement. Our guarantee requires a problem-dependent $\beta$ that approaches one as the number of components grows. In contrast, no fixed $\beta$ guarantees convergence to stationarity for every number of components. We show this by constructing a family of smooth convex problems on which, for any fixed $\beta\in[0,1)$, Muon remains bounded away from stationarity once the number of components is sufficiently large. This failure occurs for every sequence of permutations and persists even with small stepsizes. These results suggest two practical tuning rules: if the training loss stops decreasing even after reducing the stepsize, increase $\beta$; when increasing $\beta$, reduce the stepsize accordingly.
- [552] arXiv:2611.18150 (cross-list from q-fin.TR) [pdf, html, other]
-
Title: Who Should Be on the Panel? Task-Conditioned Persona Selection for LLM Trading AgentsComments: Accepted at the AI4F workshop, ICAIF 2026. Code: this https URLSubjects: Trading and Market Microstructure (q-fin.TR); Artificial Intelligence (cs.AI)
Multi-agent LLM systems are built by fixing a cast of roles and then designing how the cast coordinates. Which agents should make up the group for a given task is left to the designer, and the same cast is reused for every task. Financial trading makes the cost of that choice visible: stocks trade differently, and the trading styles suited to one need not suit another. We study task-conditioned persona selection, composing the group from the characteristics of the task it will face, and instantiate it with FinPanel, an LLM multi-agent trading system in which a recruiter selects a panel of named trader personas for each stock from a direction-blind statistical sketch of how the stock trades, each persona gives an independent daily call in its own style, and a manager commits the trade after reading the panel's arguments and their desk-kept, hierarchical track records. Because nothing is returned to the traders, the panel stays the panel that was selected, and it can trade a new stock from an empty record. On eight US stocks over the out-of-knowledge year 2025, against three prior LLM trading agents re-implemented in the same framework and against rule-based, predictive and reinforcement-learning baselines, the selected panel is ahead of a fixed panel of the same size and of the prior agents run as their papers prescribe, with a shallower drawdown than buy-and-hold on every stock; it keeps its performance with no warm-up and with no record at all, whereas the memory-augmented agents' returns swing with their warm-up; and neither more seats nor the manager's reading of the arguments substitutes for selecting the right ones. Who is in the group is a design decision to be made per task, before how the group coordinates.
- [553] arXiv:2611.18164 (cross-list from stat.ML) [pdf, html, other]
-
Title: Detection and Classification in Latent Spaces: High-Dimensional Analysis and ValidationSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Foundation models produce high-dimensional latent representations in which suitably designed statistical decision rules can be effective yet analytically tractable. We study such detectors in the high-dimensional regime, where latent dimension and training sample size grow proportionally. For linear and quadratic statistics, we derive explicit large-deviation characterizations of false-alarm and miss-detection performance under the Neyman-Pearson paradigm, and show how the binary analysis extends to multi-class classification. The resulting error probabilities and corresponding decay rates exhibit non-monotonic dependence on the shape parameter, including double-descent and, in some regimes, multiple-descent. Experiments on vision, text, and audio datasets using modern foundation models show close agreement between theoretical predictions and empirical performance. In particular, the proposed framework allows us to set a prescribed false-alarm level and accurately predict the corresponding miss-detection probability. Overall, our results provide a theoretically grounded and interpretable framework with competitive empirical performance across different modalities.
- [554] arXiv:2611.18173 (cross-list from math.OC) [pdf, html, other]
-
Title: Safety Certificates for Systems under Attack via Stackelberg Set GamesSubjects: Optimization and Control (math.OC); Systems and Control (eess.SY)
We study a Stackelberg set selection game for linear systems subject to bounded disturbances and additive attacks. The defender selects the vector parameter of a polytopic safe set template; the attacker, after observing this choice, selects the vector parameter of an admissible attack set. The payoff is a signed, facet-normalised invariance margin that tests invariance exactly and quantifies either the available contraction or the facet inflation needed when invariance fails. Both controlled and robust controlled invariance are treated, corresponding to different information patterns for the control input. Exploiting the structure of parameterised polyhedra, we show that within each validity region the margin is concave in the attack parameter, which reduces the follower minimisation to finitely many extreme parameters. The complete max--min game is then solved by a finite family of LPs, one per defender validity region, each enforcing all attacker regions under a common defender decision and guaranteed margin. The framework handles hard state constraints and additional costs related to the safe set size, and recovers optimal set parameters together with state and attack realisations that activate the certificate. A load frequency control example under false data injection illustrates the construction, the validity decomposition, and the two information patterns.
- [555] arXiv:2611.18180 (cross-list from stat.ML) [pdf, html, other]
-
Title: LaMe: a Joint Latent Mediation Framework for High-Dimensional SystemsSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Quantitative Methods (q-bio.QM); Methodology (stat.ME)
Mediation analysis seeks to understand how an exposure influences an outcome through intermediate variables. However, as increasingly rich datasets become available, hundreds to thousands of potential mediators make these systems difficult to model and hard to interpret. Latent-variable approaches offer a natural way to accommodate these high-dimensional settings. However, existing latent mediation methods face limitations, including overparameterization, biased inference, or limited scalability. We develop LaMe, a joint linear-Gaussian framework that summarizes a high-dimensional exposure-mediator-outcome system through a small number of latent variables. We identify natural direct and indirect effects under an explicit assumption that the manifest measurements fully capture the latent mediators, and characterize conditions under which the latent representation effectively restricts the observed cross-covariance structure. Simulations assess recovery of observed association parameters, natural effects, and manifest features associated with latent mediation pathways across a range of challenging settings. Four real-world applications demonstrate competitive predictive performance with compact latent representations, including single-dimensional representations of biomarker and proteomic measurements. Together, these results establish LaMe as a practical framework for estimating and summarizing multivariate mediation systems under explicit structural and measurement assumptions.
- [556] arXiv:2611.18181 (cross-list from eess.SP) [pdf, html, other]
-
Title: Linear Dynamic Range Extension for Rydberg Atomic Receiver via Band-Limited ReconstructionComments: 6 pages, 7 figuresSubjects: Signal Processing (eess.SP); Information Theory (cs.IT)
Rydberg-atom receiver offers a promising approach to radio-frequency (RF) signal reception in wireless communication and sensing systems, combining high sensitivity with the potential for broad frequency coverage. However, the limited linear dynamic range (LDR) of the atomic system causes nonlinear distortion at high input signal levels, thereby restricting the receiver's effective linear input range. To address this issue, we analyze the mechanisms underlying nonlinear distortion in Rydberg atomic receivers and propose a physics-informed band-limited iterative reconstruction (PI-BLIR) method. Specifically, by exploiting the known nonlinear atomic response and signal bandwidth prior, we formulate a constrained maximum-likelihood problem to reconstruct the distorted complex intermediate-frequency (IF) waveform. With the bandwidth constraint absorbed into the frequency-domain parameterization, the constrained maximum-likelihood problem is transformed into an unconstrained optimization problem, which is solved using gradient descent. The proposed method relies solely on digital post-processing and requires no changes to the local-oscillator (LO) field strength, optical parameters, or atomic operating conditions. Simulation results demonstrate that PI-BLIR extends the receiver's LDR by more than 14 dB and enhances the demodulation performance of the atomic receiver.
- [557] arXiv:2611.18243 (cross-list from eess.IV) [pdf, html, other]
-
Title: D2ToF: One-Step Diffusion for Depth Super-Resolution from Lightweight Single-Photon LiDAR WaveformsSubjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
Single-photon LiDAR (SPL) sensors offer exceptional temporal resolution and single-photon sensitivity for depth imaging. However, cost-effective SPL arrays typically suffer from extremely low spatial resolutions (e.g. $8\times 8$ pixels), limiting their real-world utility. In this paper, we propose a novel, one-step super-resolution framework, D2ToF, Distilled Diffusion for Time-of-Flight, that bridges the gap between high-fidelity generative modeling and real-time execution. Our approach uses an amortized variational distillation strategy that allows for a measurement-based data consistency term to be incorporated directly into the one-step training objective. This encourages that the synthesized high-resolution depth maps are consistent with the captured low-resolution Time-of-Flight (ToF) histograms. To overcome the severe ill-posedness of upsampling extremely degraded inputs, we introduce a multimodal guidance strategy. By leveraging a foundational monocular depth model to extract structural priors from a high-resolution reflectivity image, our network is guided by a rich triad of the ToF histogram, reflectivity map, and relative depth estimate. Experimental results outperform evaluated baselines, achieving $16\times$ spatial super-resolution (from $8\times 8 \times 100$ histograms to $128\times 128$ depth maps) at an accelerated inference speed of approximately $83$ milliseconds.
- [558] arXiv:2611.18307 (cross-list from physics.chem-ph) [pdf, html, other]
-
Title: AI for Parameter Selection in Sample-based Quantum DiagonalizationDoriela Grabocka, Dimitrios Christofidellis, Federico Zipoli, Anna van Elst, Patrick Ruch, Antonio Foncubierta-Rodríguez, Siddhartha Mishra, Matteo ManicaSubjects: Chemical Physics (physics.chem-ph); Artificial Intelligence (cs.AI)
Sample-based Quantum Diagonalization (SQD) methods provide a way to compute molecular ground-state energies, but their accuracy and runtime depend sensitively on configuration choices such as the subspace dimension, propagation timestep, and the allocation of a fixed sampling budget across circuits, for which no formal selection method exists. In this work, we explore whether oracle models can predict the quantum runtime and approximation error from the Hamiltonian and the candidate configuration. Using synthetic Hamiltonians with controllable spectral properties, we show that machine-learned models can map a Hamiltonian directly to the best-ranked configuration within a predefined grid for the SKQD algorithm, validated against classical Krylov propagation, achieving strong rank correlations between true and predicted error and runtime, Spearman $\rho = 0.969$ and $\rho = 0.905$, respectively. Extending the approach to 50 molecular Hamiltonians (20-50 qubits), with reference ground-state energies from classical calculations, we train regressors that predict the simulation error and runtime for a given molecule--configuration pair ahead of execution, and classifiers to distinguish, from 600 SqDRIFT quantum-hardware runs, whether a configuration achieves a ground-state energy within chemical accuracy ($1.6\times10^{-3}$ Ha $\approx 1$ kcal/mol). For the molecular Hamiltonians, the models additionally provide an a priori feasibility and resource estimate, predicting whether a molecule can reach chemical accuracy on current hardware and at what runtime, thus correctly classifying held-out molecules at the molecule level, achieving balanced accuracy 61.8% in-distribution and 100% for five larger held-out molecules.
- [559] arXiv:2611.18325 (cross-list from quant-ph) [pdf, html, other]
-
Title: Formulating Uncomputation Beyond Classical Ancilla UseSubjects: Quantum Physics (quant-ph); Programming Languages (cs.PL)
Ancilla qubits are routinely used to simplify quantum circuit construction and reduce implementation costs, but their use introduces the burden of designing proper uncomputation afterward. To alleviate this, prior work has developed automated uncomputation synthesis based on a computational-basis formulation of ancilla use. However, this formulation is tailored to classically structured ancilla computations. For more general circuits, it may be undefined, yield unintended functionality, or fail to synthesize uncomputation. It therefore falls short of providing a precise and general formulation of uncomputation beyond classical ancilla use.
To address these limitations, we propose a new formulation of ancilla use that specifies the operation of each gate through two complementary aspects: (1) the functionality it implements and (2) the transformation it induces on ancilla roles. Based on this specification, we define uncomputation as restoring the initial ancilla role while implementing the identity functionality. This perspective enables a simple Use-first strategy for automatically synthesizing uncomputation in cases both within and beyond the scope of prior approaches. We further develop a complete inference system for reasoning about specifications, together with a derivation-directed procedure for automatically synthesizing efficient uncomputation, enabling systematic uncomputation synthesis even in challenging cases beyond the scope of our automatic strategy. - [560] arXiv:2611.18330 (cross-list from eess.AS) [pdf, html, other]
-
Title: Marimba: A System One Model for SpeechSubjects: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Sound (cs.SD)
Speech applications require decisions about term presence, turn completion, and language within an audio window. We present Marimba, an interface for typed speech decisions built on frozen Qwen3-ASR 0.6B features. Lexical questions use zero-shot scoring of ASR transcripts; non-lexical questions use a question-conditioned head and a separate layer-18 turn-end readout. Language identification uses the ASR model's own language scores. On private conversational test sets, decoder AUROC is 0.80-0.92; on public read speech it is 0.92-0.97. Speech Commands four-way accuracy is 0.93; binary lexical ECE is at most 0.08. Personal-information detection, which passes the greedy transcript to a private, curated PII pipeline, reaches item-level F1 0.904 on synthetic Brazilian Portuguese speech (1.000 on reference transcripts). Layer-12 speaker readouts detect speaker change with AUROC 0.966-0.980 on AMI and VoxConverse, above an ECAPA baseline, and count speakers (0, 1, 2, or 3+) more accurately than pyannote 3.1. The turn-end readout reaches AUROC 0.930 versus Smart Turn v3.2 at 0.982 on Smart Turn's own test set. On a subset of ICSI meetings labeled by next speaker, AUROC is 0.716 versus 0.712; the paired confidence interval includes zero.
- [561] arXiv:2611.18352 (cross-list from math.OC) [pdf, html, other]
-
Title: Online Switching Control as Learning with ExpertsSubjects: Optimization and Control (math.OC); Machine Learning (cs.LG); Systems and Control (eess.SY)
This paper addresses the problem of online switching control for discrete-time nonlinear systems where a controller is sequentially selected from a finite set of candidates. The goal is to generate an adaptive sequence of controllers to minimize the cumulative cost under unknown disturbances and cost functions. Specifically, we consider a full-information feedback setting where realized disturbance effects and cost functions are available after each selection. Using counterfactual costs computed from historical data, we reduce the online switching control problem to the experts problem with switching costs. Then, we propose an adaptive switching algorithm based on the Hedge algorithm with a switching suppression mechanism. Notably, the proposed algorithm achieves an expected regret of $O(\sqrt{T\log N})$, where $T$ is the time horizon and $N$ is the number of candidate controllers, while ensuring uniform boundedness of the state. This matches the optimal regret order for the standard experts problem.
- [562] arXiv:2611.18378 (cross-list from quant-ph) [pdf, html, other]
-
Title: Structure makes quantum circuits trainableComments: 83 pages, 4 figuresSubjects: Quantum Physics (quant-ph); Machine Learning (cs.LG)
The scalability of variational quantum algorithms is limited by barren plateaus, where gradients vanish exponentially in system size. This worst-case result raises fundamental questions about the trainability of quantum AI models. We present a major advance towards resolving this trainability question by answering whether a given quantum circuit has a barren plateau. We do this using an analytical framework that bounds the gradient variance for arbitrary input states $\rho$ and observables $\mathcal{M}$ by decomposing the unitary representation of the associated compact Lie group. When the circuit segments on both sides of a gate form $2$-designs, we obtain the variance exactly and prove a sufficient condition for trainability: if $\rho^{\otimes 2}$ and $\mathcal{M}^{\otimes 2}$ are sufficiently concentrated on the same irreducible invariant subspace of polynomial dimension, and the differentiated gate moves this subspace by an inverse-polynomial amount, then the gradient variance is at least inverse-polynomial. We further present a sampling algorithm that estimates the gradient variance from runs of the ansatz, without constructing the representation-theoretic projectors. Our algorithm decides whether the variance exceeds a tolerance $\tau$ with the number of runs polynomial in $1/\tau$, including cases in which the bounds are inconclusive. This resolves the question of the trainability of the circuit before expensive training procedures.
- [563] arXiv:2611.18389 (cross-list from physics.soc-ph) [pdf, html, other]
-
Title: A systematic comparison of physics-informed neural networks and mechanistic models for epidemic modeling and forecastingSubjects: Physics and Society (physics.soc-ph); Machine Learning (cs.LG)
Physics-Informed Neural Networks (PINNs) have recently been applied to epidemic modeling and forecasting with promising results. These methods integrate mechanistic knowledge with data-driven learning. Yet, their comparative performance against well-established, and simpler epidemic models is underexplored. Here, we conduct a systematic comparison of a PINN-based framework and a stochastic mechanistic model. Both approaches share the same age-structured SEIRD core and rolling-window forecasting pipeline, in a deliberately minimal setting. We evaluate their retrospective fits and probabilistic forecasts in two synthetic scenarios and the first COVID-19 wave in London. Interestingly, while PINNs underperform in synthetic scenarios, they achieve better performance on real data. Furthermore, PINNs provide narrower uncertainty intervals and sharper estimation of parameters in synthetic scenarios. Our findings confirm PINNs as a promising tool in epidemiology. At the same time, the results reaffirm the value of simpler approaches such as traditional compartmental models, and promote additional benchmarking studies.
- [564] arXiv:2611.18394 (cross-list from stat.ML) [pdf, html, other]
-
Title: Lipschitz-Constrained Discriminators in Generative Modeling: Asymptotic Behavior and Support MismatchSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
We study the asymptotic behavior of optimal discriminators associated to $L-$Lipschitz regularized KL-divergences of two probability distributions $Q$ and $P$ that interpolate between the rescaled Wasserstein$-1$ metric by some parameter $L>0$ and the KL divergence. The parameter $L$ plays the role of a Lipschitz constraint and is viewed as a regularization parameter: small values of $L$ enforce a more geometry-aware discriminator, while as $L$ increases the geometric regularization is progressively relaxed. It is a known result [3] that such divergences converge to the KL divergence as $L\to\infty$. In this paper, we characterize the asymptotic behavior of the optimal discriminators as $L\to\infty$ in different regimes: Firstly, under $Q\ll P$ and the Radon-Nikodym derivative $\frac{dQ}{dP}$ is bounded from above and bounded away from zero. Secondly, when $Q$ and $P$ are discrete distributions satisfying $Q\ll P$, and finally when $Q$ and $P$ are non absolutely continuous discrete distributions. This result motivates training strategies which involve discriminators, such as adversarial generative models. In practice, the data distribution is often represented by a discrete empirical measure, while, especially at the early stages of the training, the model distribution may have little or no overlap with the empirical data distribution.
- [565] arXiv:2611.18420 (cross-list from cond-mat.mtrl-sci) [pdf, html, other]
-
Title: Graph diffusion transformer for interpretable generation and inverse design of disordered atomic structuresComments: 24 pages, 4 figuresSubjects: Materials Science (cond-mat.mtrl-sci); Disordered Systems and Neural Networks (cond-mat.dis-nn); Mesoscale and Nanoscale Physics (cond-mat.mes-hall); Machine Learning (cs.LG); Computational Physics (physics.comp-ph)
Atomic disorder offers a potent degree of freedom to tailor material properties, yet the resulting vast configuration space obscures structure-property relations and hinders rational inverse design. Here we introduce Graph Amorphous Diffusion Transformer (GaDiT), a generative model that combines dynamic periodic graphs, masked conditioning and an interval-weighted training-noise distribution. GaDiT generates high-quality disordered structures while accelerating sampling by two orders of magnitude compared with existing corrector-based approaches, as demonstrated across two-dimensional amorphous carbon and three-dimensional amorphous silicon and silica. Moreover, GaDiT supports conditional generation under single or combined structural features, physical properties and preparation conditions in a unified model, and enables inverse design that expands accessible structure-property ranges beyond the training distribution. To interpret these structures and the trained network, we propose a conditioned atomic response mapping analysis, revealing both the global structural descriptors and the local atomic environments that govern target properties. Overall, GaDiT establishes a general framework for interpretable generation, inverse design and structure-property discovery for disordered atomic structures.
- [566] arXiv:2611.18424 (cross-list from math.OC) [pdf, html, other]
-
Title: Balancing-based model reduction for output-constrained optimal control of switched linear systemsComments: 24 pagesSubjects: Optimization and Control (math.OC); Numerical Analysis (math.NA)
We develop a balanced-truncation (BT) framework for model-order reduction of optimal control problems governed by switched linear systems, with quadratic tracking costs, convex input constraints, and Moreau--Yosida-regularized output constraints. For general Petrov--Galerkin reduced models, we derive an a posteriori bound for the error in the optimal value function, and combine it with BT-based estimates of the dynamical-system reduction error. We show that BT preserves the duality structure underlying the first-order optimality conditions, making first-reduce-then-optimize and first-optimize-then-reduce equivalent. Consequently, reducing the forward system also reduces the associated optimality system, leading to a certified, online-efficient, and online-adaptive approximation of the parameterized solution operator. The reduced model can therefore be reused across different regularization parameters, constraint sets, switching signals, and reference trajectories.
- [567] arXiv:2611.18448 (cross-list from quant-ph) [pdf, html, other]
-
Title: Learning observable propagation with sparse local momentsComments: 33 pages, 16 figuresSubjects: Quantum Physics (quant-ph); Numerical Analysis (math.NA)
We propose a unified framework for predicting local moment dynamics in high-dimensional many-body systems from short-time trajectories. Thanks to the locality of the dynamics, we construct the transition moment matrices by projecting the Markov operator onto low-order connected-cluster basis functions supported near the target observable. Exploiting the spatial decay of time-lagged correlations, {these matrices} can be approximated in sparse-plus-low-rank form using trajectory samples, with computational cost scaling linearly in the system dimension and the number of trajectories. The learning error can be further reduced by correcting the memory effect induced by the Galerkin projection. We numerically demonstrate the effectiveness of the framework in extrapolating the dynamics of various many-body systems, including Langevin and Glauber dynamics for classical systems as well as Lindblad dynamics for open quantum systems. In addition, for a class of single-site Glauber dynamics, we establish error bounds for sparse-plus-low-rank compression and uniform-in-time bounds for connected-cluster approximation and observable propagation with compressed transition moment matrices.
- [568] arXiv:2611.18469 (cross-list from stat.ML) [pdf, html, other]
-
Title: Contrastive latent representation learning with denoising score matchingComments: 20 pages, 2 figures, 12 tables. Code: this https URLSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Contrastive representation learning methods operate by pulling different views of the same data sample together while pushing them away from views of different samples. A fundamental challenge is the estimation of the repulsive force, which typically requires large batch sizes, can exhibit high variance, and is susceptible to the curse of dimensionality. We introduce ScoreCLR, a contrastive representation learning method that offers a novel way to estimate this repulsion term. We consider the framework of representation learning by maximizing the mutual information between two latent views of a sample. By deriving the gradient of the Shannon entropy with respect to the encoder parameters, we obtain an equivalent objective that involves the score of the latent representation. We then use denoising score matching techniques to estimate this score with a lightweight network that is trained jointly with the encoder. We further propose regularization techniques using different noise levels and a Gaussian prior on the latent representation. ScoreCLR replaces the batch of explicit negative samples in contrastive methods with a learned repulsion whose cost is linear in the batch size rather than quadratic. On CIFAR-10 and ImageNet-100 it does so with minimal loss of linear-probe accuracy compared to SimCLR under a shared recipe, while producing projector outputs of higher effective rank, closer to the Gaussian prior.
- [569] arXiv:2611.18480 (cross-list from stat.ML) [pdf, html, other]
-
Title: Bregman geometry in regular exponential families: dual representations and relative smoothnessSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC)
Regular exponential-family distributions are ubiquitous in statistics due to their appealing and convenient properties. Notably, they can be expressed in terms of Bregman divergences, revealing fundamental connections between statistics, convex analysis, and information geometry. However, this geometric structure has not been systematically exploited to characterize these distributions with respect to their expectation parametrization. Building on a projective characterization, we establish a general construction of equivalent dual Bregman representations, together with normalization and uniqueness results. These representations express log-densities as ordinary Bregman divergences in dual coordinate systems, providing geometric characterizations of the parameter space. So-called $\beta$- and $\alpha$-scaled dual representations emerge as particular instances of this framework. We further derive dual Bregman representations of conjugate priors and establish that they preserve conjugacy. Applications to classical distributions illustrate the geometric structures underlying seemingly different statistical models. Moreover, we show that these representations provide natural reference geometries for establishing relative smoothness of the primal Bregman representation of the distribution, with explicit bounds that are sharp in the scalar and separable settings. An illustrative application to the Poisson distribution sheds new light on several works in the literature, providing complementary geometric justifications for reference geometries while deriving existing algorithms.
- [570] arXiv:2611.18504 (cross-list from quant-ph) [pdf, html, other]
-
Title: Agentic AI for Discovering Fault-Tolerant Quantum CircuitsComments: 20 pages, 6 figuresSubjects: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI)
Scaling quantum computing toward practical quantum advantage is currently limited by the efficiency of circuits that operate on information encoded by quantum error-correcting codes. The full stack of quantum computing encompasses several tasks: preparation of logical code states, implementation of logical gates, logical routines on encoded qubits, etc. Yet, these circuits are largely designed by hand for individual cases, with some recent works incorporating machine learning for certain aspects. We formulate this problem as agentic evolutionary program synthesis, wherein a large language model, leveraging its pretraining knowledge and tool-use proficiency, and guided by a circuit verifier, evolves a program that defines a Clifford circuit for a wide range of curated tasks such as logical state preparation, logical gate design, and logical routines such as code-switching, code growth, etc. Our approach consistently finds fault-tolerant circuits for all of these tasks, each verified by an exhaustive check over faults.
- [571] arXiv:2611.18541 (cross-list from math.RA) [pdf, html, other]
-
Title: No free lattice contains an infinite ascending chain of copies of FL(3)Comments: 59 pages, 13 figures; the main theorem is checked in Agda; the formal development is available at this https URLSubjects: Rings and Algebras (math.RA); Logic in Computer Science (cs.LO); Group Theory (math.GR); Logic (math.LO)
No free lattice contains an infinite strictly ascending chain of sublattices isomorphic to FL(3), the free lattice on three generators; this answers a question that R. A. Dean asked and the second author's 1997 notes on the problem left open. Every such chain would yield, inside FL(3), a coherent chain of proper self-embeddings $\Theta_0, \Theta_1, \Theta_2, \dots$ with $\Theta_k \circ \Theta_j = \Theta_j$ for $j < k$, so that each later embedding fixes the image of each earlier one pointwise. Every proper self-embedding of FL(3) whose fixed sublattice contains a free generating triple fixes one of the three generators, say $z$, because a power of it or of its dual lies above the identity, and a self-embedding that moves every generator and lies above the identity fixes at most eleven elements. Up to duality, a proper self-embedding fixing the image of such a $\Theta_1$ pointwise fixes only elements above $z \wedge (x \vee y)$, where no free generating triple containing $z$ can lie. We also verify the one-step example of the 1997 notes, prove that no step toward such a chain is self-similar, and show that the configuration the argument rules out at length three does occur at length two.
Without freeness the answer reverses: FL(3) contains an infinite strictly ascending chain of 3-generated sublattices, so the lattice analogue of Takahasi's chain condition for free groups fails. The theorem itself is checked by the Agda proof assistant under its safe flags, for free lattices over generators with decidable equality, and its formal proof uses no classical principle: finite searches and an induction on a finite set replace the infinite pigeonhole principle by which a chain yields a coherent chain. - [572] arXiv:2611.18542 (cross-list from stat.ML) [pdf, html, other]
-
Title: Elliptic Sampling: Optimal Interpolant Schedules for Gaussians and BeyondSubjects: Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Generative modeling based on stochastic interpolants requires specifying two key choices: an interpolant schedule bridging the initial and target distributions, and an efficient discretization of the resulting probability-flow ODE. In this work, we introduce a new framework for joint schedule/discretization design based on the analyticity properties of the (complexified) ODE trajectories, which in turn ensure good numerical behavior. Our formulation can be solved exactly via conformal mapping, yielding a natural family of variance-preserving elliptic schedules, parametrized by Jacobi elliptic functions. Pairing an elliptic schedule with a tailored discretization, we obtain Zoloflow, a new sampling algorithm closely connected to Zolotarev's optimal rational function approximants. Crucially, both the elliptic schedule and Zoloflow explicitly incorporate the condition number $\kappa$ of the target distribution and reduce to the classical trig schedule in the limit $\kappa \rightarrow 1$. For Gaussian targets, Zoloflow achieves minimax optimality among all oblivious score-span algorithms, attaining total-variation error $\varepsilon$ using $\Theta(\log\kappa\,\log(\sqrt d/\varepsilon))$ (exact) score queries. For the log-concave case, the elliptic schedule allows sampling with at most $\varepsilon$ 2-Wasserstein-error by requiring merely $\widetilde{O}(\sqrt{d} \, \mathrm{polylog}(\kappa, 1/\varepsilon))$ score evaluations. Complementing these theoretical results, we empirically evaluate the elliptic schedule on general (non-log-concave) target distributions, including detailed comparisons with alternative schedule proposals on both synthetic benchmarks and existing image generation datasets.
- [573] arXiv:2611.18561 (cross-list from q-bio.NC) [pdf, other]
-
Title: FETTUCCINE: Fast and efficient brain-to-text decoding on mobile devicesJonathan D. McCart, Pranav I. Deevi, Mehdi Azabou, Nanda H. Krishna, Caleb A. McKinney, Nicholas S. Card, Sergey D. Stavisky, Chethan PandarinathComments: NeurIPS 2026Subjects: Neurons and Cognition (q-bio.NC); Machine Learning (cs.LG)
Recent advances in speech neuroprostheses have enabled new communication avenues for people who have lost the ability to speak due to neurological illness or injury. These systems can decode neural activity during attempted speech into text by relying on language priors to achieve high decoding accuracy. This performance currently comes at the cost of using a compute-intensive language modeling stack, presenting a major bottleneck for real-world use. Further, current language models incur highly variable and disruptive inference latencies, limiting communication throughput and preventing users from reliably participating in natural conversation. We introduce FETTUCCINE, an end-to-end brain-to-text framework that leverages the speed and portability of automatic speech recognition (ASR) models to address the aforementioned practical challenges, while achieving usable (<10%) word error rates (WER), efficient adaptation to neural non-stationarities, and reliable performance on mobile devices. When evaluated using publicly available data from an intracortical speech neuroprosthesis user, FETTUCCINE outperforms previous end-to-end methods, achieving as low as 4.39% WER. Most notably, our models successfully run on a commercially-available mobile device with throughput reaching up to 370x real-time, providing the first demonstration of brain-to-text decoding on a mobile device. We also show that our models can be successfully finetuned to extend usable performance to future days. Beyond performance, our approach enables localization of the neural features most relevant for decoding via gradient-based saliency maps, which we show align with well-established physiological priors across different ASR model families. Taken together, our results show that FETTUCCINE overcomes core infrastructural and computational barriers, yielding a new class of portable brain-to-text communication.
Cross submissions (showing 60 of 60 entries)
- [574] arXiv:1809.09517 (replaced) [pdf, html, other]
-
Title: Time Relaxation with Iterative Modified Lavrentiev RegularizationSubjects: Numerical Analysis (math.NA)
We study a time relaxation regularization of the Navier--Stokes equations, \[u_t+u\cdot\nabla u+\nabla p-\nu\Delta u+\tfrac{\chi}{\delta}(u-D_N\bar u)=f,\] in which $\bar u$ is the differential filter of radius $\delta$ applied to $u$ and $D_N$ is the iterated modified Lavrentiev deconvolution operator with regularization parameter $\alpha$. We first show that the relaxation term has a simple closed form: it is a filtered hyperviscosity, $\nu_N(-\Delta)^{N+1}G_{\sqrt\alpha\delta}^{N+1}u$ with $\nu_N=\chi\alpha^{N+1}\delta^{2N+1}$, where $G_{\sqrt\alpha\delta}$ is the differential filter of the smaller radius $\sqrt\alpha\,\delta$. From this we obtain sharp consistency bounds, the existence of weak solutions with an energy inequality, and an explicit estimate showing that the energy in scales below any given length is $O(\delta/\chi)$ as $\chi\to\infty$. Extending a result of Pakzad for van Cittert deconvolution, we bound the time-averaged energy dissipation rate by $\frac{U^3}{\ell}\big(1+\Rey^{-1}+\Rey_N^{-1}\big)$, where $\Rey_N=UL^{2N+1}/\nu_N$ is the relaxation Reynolds number. We then develop a K41-type similarity theory. The model's micro-scale solves an explicit scalar equation, which has no solution below a threshold value of $\chi$ and two solutions above it; the larger one is the cut-off scale. Requiring the micro-scale to equal the filter radius gives the scaling $\chi\simeq U(\delta/L)^{1/3}(1+1/\alpha)^{N+1}$, for which the consistency error of the relaxation term is $O(\delta^{2N+4/3})$ uniformly in $\alpha$.
- [575] arXiv:2101.00834 (replaced) [pdf, html, other]
-
Title: Symbolic Control for Stochastic Systems via Finite Parity GamesJournal-ref: Majumdar, R., Mallik, K., Schmuck, A.K. and Soudjani, S., 2024. Symbolic control for stochastic systems via finite parity games. Nonlinear Analysis: Hybrid Systems, 51, p.101430Subjects: Systems and Control (eess.SY); Logic in Computer Science (cs.LO)
We consider the problem of computing the maximal probability of satisfying an omega-regular specification for stochastic nonlinear systems evolving in discrete time. The problem reduces, after automata-theoretic constructions, to finding the maximal probability of satisfying a parity condition on a (possibly hybrid) state space. While characterizing the exact satisfaction probability is open, we show that a lower bound on this probability can be obtained by (I) computing an under-approximation of the qualitative winning region, i.e., states from which the parity condition can be enforced almost surely, and (II) computing the maximal probability of reaching this qualitative winning region. The heart of our approach is a technique to symbolically compute the under-approximation of the qualitative winning region in step (I) via a finite-state abstraction of the original system as a 2.5-player parity game. Our abstraction procedure uses only the support of the probabilistic evolution; it does not use precise numerical transition probabilities. We prove that the winning set in the abstract 2.5-player game induces an under-approximation of the qualitative winning region in the original synthesis problem, along with a policy to solve it. By combining these contributions with (a) a symbolic fixpoint algorithm to solve 2.5-player games and (b) existing techniques for reachability policy synthesis in stochastic nonlinear systems, we get an abstraction-based algorithm for finding a lower bound on the maximal satisfaction probability. We have implemented the abstraction-based algorithm in Mascot-SDS (Majumdar et al., 2020), where we combined the outlined abstraction step with our recent tool FairSyn. We evaluated our implementation on the nonlinear model of a perturbed bistable switch from the literature. We outperform a recently proposed tool for solving this problem by a large margin.
- [576] arXiv:2108.11569 (replaced) [pdf, html, other]
-
Title: Robust Long-Tailed Learning under Label NoiseComments: Paper published on Frontiers of Computer Science (FCS 2026)Subjects: Machine Learning (cs.LG)
Long-tailed learning aims to enhance the generalization performance of underrepresented tail classes. However, previous methods have largely overlooked the prevalence of noisy labels in training data. In this paper, we address the challenge of noisy labels in long-tailed learning. We identify a critical issue: the commonly used small-loss noisy label detection criterion fails to perform effectively in long-tailed class distributions. This failure arises from the inherent bias of deep neural networks, which tend to misclassify tail class examples as head classes, leading to unreliable loss calculations. To mitigate this, we propose a novel small-distance criterion that leverages the robustness of learned representations, enabling more accurate identification of correctly-labeled examples across both head and tail classes. Additionally, to improve training for tail classes, we replace discrete pseudo-labels with label distributions for examples flagged as noisy, resulting in significant performance gains. Based on these contributions, we introduce the Robust Long-Tail learning framework, designed to train models that are resilient to both class imbalance and noisy labels. Extensive experiments on benchmark and real-world datasets demonstrate that our approach outperforms previous methods, offering substantial performance improvements.
- [577] arXiv:2209.07351 (replaced) [pdf, html, other]
-
Title: Rethinking Round-Trip Translation for Machine Translation EvaluationComments: Accepted to Findings of ACL 2023Subjects: Computation and Language (cs.CL)
Automatic evaluation on low-resource language translation suffers from a deficiency of parallel corpora. Round-trip translation could be served as a clever and straightforward technique to alleviate the requirement of the parallel evaluation corpus. However, there was an observation of obscure correlations between the evaluation scores by forward and round-trip translations in the era of statistical machine translation (SMT). In this paper, we report the surprising finding that round-trip translation can be used for automatic evaluation without the references. Firstly, our revisit on the round-trip translation in SMT evaluation unveils that its long-standing misunderstanding is essentially caused by copying mechanism. After removing copying mechanism in SMT, round-trip translation scores can appropriately reflect the forward translation performance. Then, we demonstrate the rectification is overdue as round-trip translation could benefit multiple machine translation evaluation tasks. To be more specific, round-trip translation could be used i) to predict corresponding forward translation scores; ii) to improve the performance of the recently advanced quality estimation model; and iii) to identify adversarial competitors in shared tasks via cross-system verification.
- [578] arXiv:2301.12868 (replaced) [pdf, html, other]
-
Title: On Robustness of Prompt-based Semantic Parsing with Large Pre-trained Language Model: An Empirical Study on CodexTerry Yue Zhuo, Zhuang Li, Yujin Huang, Fatemeh Shiri, Weiqing Wang, Gholamreza Haffari, Yuan-Fang LiComments: Accepted at EACL2023 (main)Subjects: Computation and Language (cs.CL)
Semantic parsing is a technique aimed at constructing a structured representation of the meaning of a natural-language question. Recent advancements in few-shot language models trained on code have demonstrated superior performance in generating these representations compared to traditional unimodal language models, which are trained on downstream tasks. Despite these advancements, existing fine-tuned neural semantic parsers are susceptible to adversarial attacks on natural-language inputs. While it has been established that the robustness of smaller semantic parsers can be enhanced through adversarial training, this approach is not feasible for large language models in real-world scenarios, as it requires both substantial computational resources and expensive human annotation on in-domain semantic parsing data. This paper presents the first empirical study on the adversarial robustness of a large prompt-based language model of code, \codex. Our results demonstrate that the state-of-the-art (SOTA) code-language models are vulnerable to carefully crafted adversarial examples. To address this challenge, we propose methods for improving robustness without the need for significant amounts of labeled data or heavy computational resources.
- [579] arXiv:2309.08674 (replaced) [pdf, html, other]
-
Title: Fake News Detectors are Biased against Texts Generated by Large Language ModelsComments: The first two authors contributed equallySubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
The spread of fake news has emerged as a critical challenge, undermining trust and posing threats to society. In the era of Large Language Models (LLMs), the capability to generate believable fake content has intensified these concerns. In this study, we present a novel paradigm to evaluate fake news detectors in scenarios involving both human-written and LLM-generated misinformation. Intriguingly, our findings reveal a significant bias in many existing detectors: they are more prone to flagging LLM-generated content as fake news while often misclassifying human-written fake news as genuine. This unexpected bias appears to arise from distinct linguistic patterns inherent to LLM outputs. To address this, we introduce a mitigation strategy that leverages adversarial training with LLM-paraphrased genuine news. The resulting model yielded marked improvements in detection accuracy for both human and LLM-generated news. To further catalyze research in this domain, we release two comprehensive datasets, \texttt{GossipCop++} and \texttt{PolitiFact++}, thus amalgamating human-validated articles with LLM-generated fake and real news.
- [580] arXiv:2401.00788 (replaced) [pdf, html, other]
-
Title: Astraios: Parameter-Efficient Instruction Tuning Code Large Language ModelsTerry Yue Zhuo, Armel Zebaze, Nitchakarn Suppattarachai, Leandro von Werra, Harm de Vries, Qian Liu, Niklas MuennighoffComments: 25 pages (12 main), 19 figures, 8 tablesSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
The high cost of full-parameter fine-tuning (FFT) of Large Language Models (LLMs) has led to a series of parameter-efficient fine-tuning (PEFT) methods. However, it remains unclear which methods provide the best cost-performance trade-off at different model scales. We introduce Astraios, a suite of 28 instruction-tuned OctoCoder models using 7 tuning methods and 4 model sizes up to 16 billion parameters. Through investigations across 5 tasks and 8 different datasets encompassing both code comprehension and code generation tasks, we find that FFT generally leads to the best downstream performance across all scales, and PEFT methods differ significantly in their efficacy based on the model scale. LoRA usually offers the most favorable trade-off between cost and performance. Further investigation into the effects of these methods on both model robustness and code security reveals that larger models tend to demonstrate reduced robustness and less security. At last, we explore the relationships among updated parameters, cross-entropy loss, and task performance. We find that the tuning effectiveness observed in small models generalizes well to larger models, and the validation loss in instruction tuning can be a reliable indicator of overall downstream performance.
- [581] arXiv:2401.13712 (replaced) [pdf, html, other]
-
Title: Engineering Yeast Cells to Facilitate Information ExchangeNikolaos Ntetsikas, Styliana Kyriakoudi, Antonis Kirmizis, Bige Deniz Unluturk, Andreas Pitsillides, Ian F. Akyildiz, Marios LestasComments: 18 pages, 9 figures (2 of which are not colored) all .png, recently accepted for publication at TMBMCSubjects: Emerging Technologies (cs.ET); Information Theory (cs.IT); Molecular Networks (q-bio.MN)
Although continuous advances in theoretical modelling of Molecular Communications (MC) are observed, there is still an insuperable gap between theory and experimental testbeds, especially at the microscale. In this paper, the development of the first testbed incorporating engineered yeast cells is reported. Different from the existing literature, eukaryotic yeast cells are considered for both the sender and the receiver, with {\alpha}-factor molecules facilitating the information transfer. The use of such cells is motivated mainly by the well understood biological mechanism of yeast mating, together with their genetic amenability. In addition, recent advances in yeast biosensing establish yeast as a suitable detector and a neat interface to in-body sensor networks. The system under consideration is presented first, and the mathematical models of the underlying biological processes leading to an end-to-end (E2E) system are given. The experimental setup is then described and used to obtain experimental results which validate the developed mathematical models. Beyond that, the ability of the system to effectively generate output pulses in response to repeated stimuli is demonstrated, reporting one event per two hours. However, fast RNA fluctuations indicate cell responses in less than three minutes, demonstrating the potential for much higher rates in the future.
- [582] arXiv:2403.06406 (replaced) [pdf, html, other]
-
Title: When No-Reference Image Quality Models Meet MAP Estimation in Diffusion LatentsComments: Accepted by IEEE TIPSubjects: Computer Vision and Pattern Recognition (cs.CV)
Contemporary no-reference image quality assessment (NR-IQA) models can effectively quantify perceived image quality, often achieving strong correlations with human perceptual scores on standard IQA benchmarks. Yet, limited efforts have been devoted to treating NR-IQA models as natural image priors for real-world image enhancement, and consequently comparing them from a perceptual optimization standpoint. In this work, we show -- for the first time -- that NR-IQA models can be plugged into the maximum a posteriori (MAP) estimation framework for image enhancement. This is achieved by performing gradient ascent in the diffusion latent space rather than in the raw pixel domain, leveraging a pretrained differentiable and bijective diffusion process. Likely, different NR-IQA models lead to different enhanced outputs, which in turn provides a new computational means of comparing them. Unlike conventional correlation-based measures, our comparison method offers complementary insights into the respective strengths and weaknesses of the competing NR-IQA models in perceptual optimization scenarios. Additionally, we aim to improve the best-performing NR-IQA model in diffusion latent MAP estimation by incorporating the advantages of other top-performing methods. The resulting model delivers noticeably better results in enhancing real-world images afflicted by unknown and complex distortions, all preserving a high degree of image fidelity.
- [583] arXiv:2404.04369 (replaced) [pdf, html, other]
-
Title: A Fine-Grained Classification of Subquadratic Patterns for Subgraph Listing and FriendsComments: v2: Expanded version establishing a classification of all graphs with subw<2Subjects: Data Structures and Algorithms (cs.DS)
In an $m$-edge host graph $G$, all triangles can be listed in time $O(m^{1.5})$ [Itai, Rodeh '78], and all $k$-cycles can be listed in time $O(m^{2-1/{\lceil k/2 \rceil}} + t)$ where $t$ is the output size [Alon, Yuster, Zwick '97]. These classic results also hold for the colored problem variant, where the nodes of the host graph $G$ are colored by nodes in the pattern graph $H$, and we are only interested in subgraphs of $G$ that are isomorphic to the pattern $H$ and respect the colors. We study the problem of listing all $H$-subgraphs in the colored setting, for fixed pattern graphs $H$.
As our main result, we determine all pattern graphs $H$ such that all $H$-subgraphs can be listed in subquadratic time $O(m^{2-\varepsilon} + t)$, where $t$ is the output size. Moreover, for each such subquadratic pattern $H$ we determine the smallest exponent $c(H)$ such that all $H$-subgraphs can be listed in time $O(m^{c(H)} + t)$. This is a vast generalization of the classic results on triangles and cycles.
We also show the same results for two related problems: finding an $H$-subgraph of minimum total edge-weight in time $O(m^{c(H)})$, and enumerating all $H$-subgraphs in $O(m^{c(H)})$ preprocessing time and constant delay. Again we determine all pattern graphs $H$ that have complexity $c(H) < 2$, and for each such subquadratic pattern we determine the optimal complexity $c(H)$.
To prove these results, we design new algorithms and prove conditional lower bounds based on standard hypotheses from fine-grained complexity theory. For our algorithms we introduce a new ingredient that we call hyper-degree splitting, where we split tuples of nodes into high degree and low degree depending on their number of common neighbors. This technique generalizes the classic high-degree-low-degree approach (which only splits along degrees of single nodes) and is crucial to obtain optimal algorithms. - [584] arXiv:2404.16787 (replaced) [pdf, html, other]
-
Title: Enhancing Quality of Experience in Future Telecommunication Networks: A Review of Frameworks and Machine Learning AlgorithmsComments: 22 pages, 13 figures, 11 tablesSubjects: Networking and Internet Architecture (cs.NI)
Quality of experience (QoE) has emerged as a key measure of user satisfaction for multimedia services, including adaptive video streaming, immersive media (AR/VR) and cloud gaming. As services grow more varied, QoE assessment has to connect measurement with the models that act on it. Existing surveys treat the two as separate catalogues and do not classify how they are coupled, so the choice of an integrated architecture is neither replicable nor transferable. This survey addresses four questions: (1) how frameworks are coupled with ML models, (2) which QoE actions each pattern supports, (3) which patterns appear in each service domain, and (4) which problems remain open. A taxonomy of loose, tight, and distributed/federated coupling answers them, in which each work is coded by where model inference runs. This paper also identifies three enhancement mechanisms (predictive optimization, adaptive resource allocation, content-aware enhancement). Because reported accuracies use different metrics and datasets, they cannot be ranked directly. A model-selection guide is therefore derived from verified results. The study traces three generations, from static measurement through reactive ML to proactive and generative methods, including large language model integration, and examines domain constraints for immersive media, cloud gaming and the Internet of Things. Open challenges include generalizability, interpretability, model drift and privacy, while objective tuning, federated scheduling, multi-modal and mobile-edge integration are emerging directions.
- [585] arXiv:2406.01999 (replaced) [pdf, html, other]
-
Title: Random Abstract Cell ComplexesComments: 9 pages, 7 figures (plus appendix). For evaluation code, see this https URLSubjects: Data Structures and Algorithms (cs.DS); Discrete Mathematics (cs.DM); Social and Information Networks (cs.SI); Algebraic Topology (math.AT)
In machine learning and signal processing, methods based on (abstract) cell complexes (CCs) are being applied successfully in different applications. However, there is currently a lack of random models, which are necessary for a thorough evaluation of these methods, both by generating synthetic data and as null models to compare results to. We fill this gap by defining a random model for abstract cell complexes based on independent inclusion probabilities, similar to the Erdős-Rényi model for graphs. We also discuss how modifications to bias the model to incorporate more structure can be implemented. The model can also be used to lift a graph (or other cell complex) to a cell complex of arbitrary dimension. The main challenge in this is that the number of cycles (possible 2-cells) on graphs is exponential, and the same is true in higher dimensions. We overcome this complexity with an approximate sampling algorithm limited to 2-cells (polygons) that is based on a recent advance in sampling cycles on a graph. Using this algorithm, we analyze the properties of random 2-dimensional CCs and showcase its usage as a performance baseline for both topological signal processing and topological deep learning. The algorithm is available in the package py-raccoon on the Python Package Index.
- [586] arXiv:2406.15877 (replaced) [pdf, html, other]
-
Title: BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex InstructionsTerry Yue Zhuo, Minh Chien Vu, Jenny Chim, Han Hu, Wenhao Yu, Ratnadira Widyasari, Imam Nur Bani Yusuf, Haolan Zhan, Junda He, Indraneil Paul, Simon Brunner, Chen Gong, Thong Hoang, Armel Randy Zebaze, Xiaoheng Hong, Wen-Ding Li, Jean Kaddour, Ming Xu, Zhihan Zhang, Prateek Yadav, Naman Jain, Alex Gu, Zhoujun Cheng, Jiawei Liu, Qian Liu, Zijian Wang, Binyuan Hui, Niklas Muennighoff, David Lo, Daniel Fried, Xiaoning Du, Harm de Vries, Leandro Von WerraComments: Accpeted at ICLR 2025 (Oral), built with love by the BigCode community :)Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Task automation has been greatly empowered by the recent advances in Large Language Models (LLMs) via Python code, where the tasks ranging from software engineering development to general-purpose reasoning. While current benchmarks have shown that LLMs can solve tasks using programs like human developers, the majority of their evaluations are limited to short and self-contained algorithmic tasks or standalone function calls. Solving challenging and practical tasks requires the capability of utilizing diverse function calls as tools to efficiently implement functionalities like data analysis and web development. In addition, using multiple tools to solve a task needs compositional reasoning by accurately understanding complex instructions. Fulfilling both of these characteristics can pose a great challenge for this http URL assess how well LLMs can solve challenging and practical tasks via programs, we introduce BigCodeBench, a benchmark that challenges LLMs to invoke multiple function calls as tools from 139 libraries and 7 domains for 1,140 fine-grained tasks. To evaluate LLMs rigorously, each task encompasses 5.6 test cases with an average branch coverage of 99%. In addition, we propose a natural-language-oriented variant of BigCodeBench, BigCodeBench-Instruct, that automatically transforms the original docstrings into short instructions only with essential information. Our extensive evaluation of 60 LLMs shows that LLMs are not yet capable of following complex instructions to use function calls precisely, with scores up to 60%, significantly lower than the human performance of 97%. The results underscore the need for further advancements in this area.
- [587] arXiv:2410.06191 (replaced) [pdf, other]
-
Title: Benign Overfitting for Regression with Trained Two-Layer ReLU NetworksComments: There were fatal errors in the proofs. For the updated version, see https://arxiv.org/abs/2505.11621Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
We study the least-square regression problem with a two-layer fully-connected neural network, with ReLU activation function, trained by gradient flow. Our first result is a generalization result, that requires no assumptions on the underlying regression function or the noise other than that they are bounded. We operate in the neural tangent kernel regime, and our generalization result is developed via a decomposition of the excess risk into estimation and approximation errors, viewing gradient flow as an implicit regularizer. This decomposition in the context of neural networks is a novel perspective of gradient descent, and helps us avoid uniform convergence traps. In this work, we also establish that under the same setting, the trained network overfits to the data. Together, these results, establishes the first result on benign overfitting for finite-width ReLU networks for arbitrary regression functions.
- [588] arXiv:2410.13013 (replaced) [pdf, html, other]
-
Title: LEGAL-UQA: Testing Constitutional Question-Answering Systems for Urdu Speakers with a Parallel English-Urdu BenchmarkComments: 8 pagesSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Legal information assistants could help Urdu speakers understand their constitutional rights, but building reliable systems requires test data and evaluation methods suited to the language. We present LEGAL-UQA, a parallel English-Urdu benchmark of 619 question-answer pairs linked to passages from Pakistan's Constitution. Each example includes the question, reference answer, and source passage in both languages. We evaluate two components separately: answering Urdu questions when the source passage is supplied, and retrieving the passage from which each question was generated. We also compare JEV, a decision model used as an automated judge, with human grades. The findings show that model rankings depend on how answer similarity is measured, while a simple word-matching method retrieves source passages more effectively than the embedding methods tested. At the evaluated cutoff, JEV agrees with human grades less often than a baseline that gives every answer full credit. An audit also identifies cases where generated references omit information supported by the source. Together, these findings highlight the need to test retrieval and answering separately, include simple retrieval baselines, and validate automated judges and reference answers with human review when engineering legal QA software for Urdu speakers.
- [589] arXiv:2412.04236 (replaced) [pdf, html, other]
-
Title: Mapping Seven Decades of Philosophy in Colombia: Dynamic Topic Modelling of Ideas y ValoresComments: Accepted for publication at ScientometricsSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Digital Libraries (cs.DL)
Data-driven approaches to philosophy have emerged as a valuable tool for studying the history of the discipline. However, most studies in this area have focused on a limited number of journals from specific regions and subfields. We expand the scope of this research by applying dynamic topic modeling techniques to explore the history of philosophy in Colombia and Latin America. Our study examines the Colombian philosophy journal Ideas y Valores, founded in 1951 and currently one of the most influential academic philosophy journals in the region. By analyzing the evolution of topics across the journal's history, we identify various trends and specific dynamics in philosophical discourse within the Colombian and Latin American context. Our findings reveal that the most prominent topics are value theory (including ethics, political philosophy, and aesthetics), epistemology, and the philosophy of science. We also trace the evolution of articles focused on interpreting a specific philosopher's work rather than proposing new positions, and we note a salient emphasis on German philosophers such as Kant, Husserl, and Hegel across various topics throughout the journal's lifetime. Given the journal's founding aspiration toward more original, non-exegetical philosophy, we investigate whether exegetical topics became comparatively less prominent over time. Our analysis suggests no significant decline in such topics. Finally, we propose ideas for extending this research to other Latin American journals and suggest improvements for natural language processing workflows in non-English languages.
- [590] arXiv:2501.12764 (replaced) [pdf, html, other]
-
Title: Grid-based Submap Joining: An Efficient Algorithm for Simultaneously Optimizing Global Occupancy Map and Local Submap FramesComments: Accepted by IROS 2024, the code is made availableSubjects: Robotics (cs.RO)
Optimizing robot poses and the map simultaneously has been shown to provide more accurate SLAM results. However, for non-feature based SLAM approaches, directly optimizing all the robot poses and the whole map will greatly increase the computational cost, making SLAM problems difficult to solve in large-scale environments. To solve the 2D non-feature based SLAM problem in large-scale environments more accurately and efficiently, we propose the grid-based submap joining method. Specifically, we first formulate the 2D grid-based submap joining problem as a non-linear least squares (NLLS) form to optimize the global occupancy map and local submap frames simultaneously. We then prove that in solving the NLLS problem using Gauss-Newton (GN) method, the increments of the poses in each iteration are independent of the occupancy values of the global occupancy map. Based on this property, we propose a poseonly GN algorithm equivalent to full GN method to solve the NLLS problem. The proposed submap joining algorithm is very efficient due to the independent property and the pose-only solution. Evaluations using simulations and publicly available practical 2D laser datasets confirm the outperformance of our proposed method compared to the state-of-the-art methods in terms of efficiency and accuracy, as well as the ability to solve the grid-based SLAM problem in very large-scale environments.
- [591] arXiv:2502.02220 (replaced) [pdf, other]
-
Title: On the Existential Theory of the Reals Enriched with Integer Powers of a Computable NumberSubjects: Logic in Computer Science (cs.LO)
This paper investigates $\exists\mathbb{R}(r^{\mathbb{Z}})$, that is the extension of the existential theory of the reals by an additional unary predicate $r^{\mathbb{Z}}$ for the integer powers of a fixed computable real number $r > 0$. If all we have access to is a Turing machine computing $r$, it is not possible to decide whether an input formula from this theory satisfiable. However, we show an algorithm to decide this problem when: 1. $r$ is known to be transcendental, or 2. $r$ is a root of some given integer polynomial (that is, $r$ is algebraic). In other words, knowing the algebraicity of $r$ suffices to circumvent undecidability. Furthermore, we establish complexity results under the proviso that $r$ enjoys what we call a polynomial root barrier. Using this notion, we show that the satisfiability problem of $\exists\mathbb{R}(r^{\mathbb{Z}})$ is 1. in NEXPTIME if $r$ is a natural number, 2. in EXPSPACE if $r$ is an algebraic number, and 3. in 3EXP if $r$ belongs to a family of transcendental numbers including $Ï€$ and Euler's $e$. As a by-product of our results, we are able to remove the appeal to Schanuel's conjecture from the proof of decidability of the entropic risk threshold problem for stochastic games with rational probabilities, rewards and threshold [Baier et al., MFCS'23]: when the base of the entropic risk is Euler's $e$ and the aversion factor is a fixed algebraic number, the problem is in EXP.
- [592] arXiv:2502.03312 (replaced) [pdf, html, other]
-
Title: An 'Experimental Mathematics' Approach to Stolarsky Interspersions via Automata TheoryComments: This revision fixes up numerous typographical errors and adds some new resultsSubjects: Formal Languages and Automata Theory (cs.FL); Discrete Mathematics (cs.DM); Number Theory (math.NT)
We look at the Stolarsky interspersions (such as the Wythoff array) one more time, this time using tools from automata theory. These tools allow easy verification of many of the published results on these arrays, as well as proofs of new results.
- [593] arXiv:2502.08739 (replaced) [pdf, html, other]
-
Title: A Taxonomy of Real Faults in Hybrid Quantum-Classical ArchitecturesComments: ACM Transactions on Software Engineering and Methodology (TOSEM)Subjects: Software Engineering (cs.SE)
With the popularity of Hybrid Quantum-Classical architectures, particularly noisy intermediate-scale quantum (NISQ) architectures, comes the need for quality assurance methods tailored to their specific faults. In this study, we propose a taxonomy of faults in Hybrid Quantum-Classical architectures accompanied by a dataset of real faults in the identified categories. To achieve this, we empirically analysed open-source repositories for fixed faults. We analysed over 5000 closed issues on GitHub and pre-selected 529 of them based on rigorously defined inclusion criteria. We selected 133 faults that we labelled around symptoms and the origin of the faults. We cross-validated the classification and labels assigned to every fault between two of the authors. As a result, we introduced a taxonomy of real faults in Hybrid Quantum-Classical architectures. Subsequently, we validated the taxonomy through interviews conducted with eleven developers. The taxonomy was dynamically updated throughout the cross-validation and interview processes. The final version was validated and discussed through surveys conducted with an independent group of domain experts to ensure its relevance and to gain further insights.
- [594] arXiv:2502.20658 (replaced) [pdf, other]
-
Title: Curating Safe Spaces for Emotion Disclosure during Illness Recovery on Video-Sharing PlatformsJiaying "Lizzy" Liu, Yunlong Wang, Allen Jue, Yao Lyu, Yiheng Su, Shuo Niu, Nathan TeBlunthuis, Yan ZhangSubjects: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY); Social and Information Networks (cs.SI)
The advent of video-sharing platforms has transformed self-expression, transcending entertainment to encompass sensitive disclosures and emotional revelations. In this submission, we discuss our study on emotional disclosure by individuals with Schizophrenia through YouTube vlogs and its implications for designing safe online spaces.
- [595] arXiv:2503.04759 (replaced) [pdf, html, other]
-
Title: Games with $ω$-Automatic Preference RelationsComments: Extended version of a MFCS 2025 paperSubjects: Logic in Computer Science (cs.LO); Computer Science and Game Theory (cs.GT)
This paper investigates Nash equilibria (NEs) in multi-player turn-based games on graphs, where player preferences are modeled as $\omega$-automatic relations via deterministic parity automata. Unlike much of the existing literature, which focuses on specific reward functions, our results apply to any preference relation definable by an $\omega$-automatic relation. We analyze the computational complexity of determining the existence of an NE (possibly under some constraints), verifying whether a given strategy profile forms an NE, and checking whether a specific outcome can be realized by an NE. When a (constrained) NE exists, we show that there always exists one with finite-memory strategies. Finally, we explore fundamental properties of $\omega$-automatic relations and their implications for the existence of equilibria.
- [596] arXiv:2503.22821 (replaced) [pdf, html, other]
-
Title: Identifying and Mitigating API Misuse in Large Language ModelsComments: Accepted at IEEE Transactions on Software Engineering (TSE)Subjects: Software Engineering (cs.SE)
API misuse in code generated by large language models (LLMs) presents a serious and growing challenge in software development, as although LLMs demonstrate impressive code generation capabilities, their interactions with complex library APIs are often error-prone and can lead to software failures and vulnerabilities. In this paper, we conduct a large-scale study of API misuse patterns in LLM-generated code by analyzing both method selection and parameter usage across Python and Java, using three representative LLMs: StarCoder-7B, Qwen2.5-Coder-7B, and GitHub Copilot. Based on extensive manual annotation of 3,209 method-level and 3,492 parameter-level misuses, we identify and categorize four recurring misuse types by building on and refining prior API misuse taxonomies. Our evaluation of the three LLMs reveals persistent challenges in API usage, particularly hallucination and intent misalignment. To address these issues, we propose this http URL, an LLM-based automatic repair approach guided by our taxonomy, which improves repair accuracy compared to baseline prompting and existing repair methods, achieving gains of up to 38.4 BLEU and up to 72 more exact matches on benchmark datasets. This work offers important insights into the current limitations of LLMs in API usage and points to directions for improving automated misuse repair in code generation systems.
- [597] arXiv:2504.16150 (replaced) [pdf, html, other]
-
Title: Evaluating the Generalizability and Explainability of Firn Image Classification with Topological FeaturesComments: Extended version with additional experiments on generalizability and explainabilitySubjects: Computer Vision and Pattern Recognition (cs.CV); Algebraic Topology (math.AT)
In this paper we evaluate the performance of topological features for generalizable and explainable classification of firn image data, with the broader goal of understanding the advantages, pitfalls, and trade-offs in topological featurization. Firn refers to layers of granular snow within glaciers that haven't yet been compressed into ice. This compaction process imposes distinct topological and geometric structure on firn that varies with depth within the firn column. However, the limited and highly correlated training data available results in naive classifiers (such as convolutional neural networks) overfitting to orientation-dependent features. This motivates us to evaluate topological data analysis (TDA) as a way to design features which capture the shape of the firn data while retaining a large degree of invariance. We evaluate two classes of topological features, sublevel set features and distance transform features, together with persistence curves, to predict sample depth from firn images. A range of challenging training-test scenarios reveals that no one choice of method dominates in all categories. In particular, we see a tradeoff between accuracy on the one hand, and generalizability and explainability (via feature importance) on the other.
- [598] arXiv:2505.11294 (replaced) [pdf, html, other]
-
Title: Bidirectional Information Flow (BIF) - A Sample Efficient Hierarchical Gaussian Process for Bayesian OptimizationComments: NeurIPS 2026 Camera ReadySubjects: Machine Learning (cs.LG)
Hierarchical Gaussian Process (H-GP) models divide problems into different subtasks, allowing for different components to address each part, making them well-suited for problems with inherent compositional structure. However, existing H-GP frameworks typically employ one-way information sharing - either top-down or bottom-up - which limits sample efficiency and slows convergence. We propose Bidirectional Information Flow (BIF), which establishes continuous two-way communication. BIF retains the modular structure of hierarchical models - the parent conditions its own posterior on child summaries, treating them as structured priors - while introducing top-down feedback to softly decompose environment observations from the parent into sub-responses. This mutual exchange improves sample efficiency, enables robust training, and allows modular reuse of learned subtask models. We prove analytically the regret of a GP with a learned kernel scales linearly with the mismatch to the true kernel, tightening in the hierarchical case to the sum of child-level errors. Ablation shows removing the downward pathway collapses child $R^2$ by up to 58%. Across synthetic, neurostimulation, and HPO benchmarks, BIF achieves up to $4\times$ higher parent $R^2$ and roughly 100% AUC improvement over vanilla GPBO, and outperforms or is the runner-up in 38/40 tests compared to the state-of-the-art baselines while supporting modular child transfer to novel composite tasks.
- [599] arXiv:2505.12358 (replaced) [pdf, html, other]
-
Title: AbFlowNet: Optimizing Antibody-Antigen Binding Energy via Diffusion-GFlowNet FusionAbrar Rahman Abir, Haz Sameen Shahgir, Md Rownok Zahan Ratul, Md Toki Tahmid, Greg Ver Steeg, Yue DongComments: Accepted at NeurIPS 2026Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Complementarity Determining Regions (CDRs) are critical segments of an antibody that facilitate binding to specific antigens. Current computational methods for CDR design utilize reconstruction losses and do not jointly optimize binding energy, a crucial metric for antibody efficacy. Rather, binding energy optimization is done through computationally expensive Online Reinforcement Learning (RL) pipelines rely heavily on unreliable binding energy estimators. In this paper, we propose AbFlowNet, a novel generative framework that integrates GFlowNet with Diffusion models. By framing each diffusion step as a state in the GFlowNet framework, AbFlowNet jointly optimizes standard diffusion losses and binding energy by directly incorporating energy signals into the training process, thereby unifying diffusion and reward optimization in a single procedure. Experimental results show that AbFlowNet outperforms the base diffusion model by 3.06% in amino acid recovery, 20.40% in geometric reconstruction (RMSD), and 3.60% in binding energy improvement ratio. ABFlowNet also decreases Top-1 total energy and binding energy errors by 24.8% and 38.1% without pseudo-labeling the test dataset or using computationally expensive online RL regimes.
- [600] arXiv:2505.18934 (replaced) [pdf, html, other]
-
Title: Chi-Square Wavelet Graph Neural Networks for Heterogeneous Graph Anomaly DetectionXiping Li, Xiangyu Dong, Xingyi Zhang, Kun Xie, Yuanhao Feng, Bo Wang, Guilin Li, Wuxiong Zeng, Xiujun Shu, Sibo WangComments: Added references on multi-view clusteringSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Social and Information Networks (cs.SI)
Graph Anomaly Detection (GAD) in heterogeneous networks presents unique challenges due to node and edge heterogeneity. Existing Graph Neural Network (GNN) methods primarily focus on homogeneous GAD and thus fail to address three key issues: (C1) Capturing abnormal signal and rich semantics across diverse meta-paths; (C2) Retaining high-frequency content in HIN dimension alignment; and (C3) Learning effectively from difficult anomaly samples with class imbalance. To overcome these, we propose ChiGAD, a spectral GNN framework based on a novel Chi-Square filter, inspired by the wavelet effectiveness in diverse domains. Specifically, ChiGAD consists of: (1) Multi-Graph Chi-Square Filter, which captures anomalous information via applying dedicated Chi-Square filters to each meta-path graph; (2) Interactive Meta-Graph Convolution, which aligns features while preserving high-frequency information and incorporates heterogeneous messages by a unified Chi-Square Filter; and (3) Contribution-Informed Cross-Entropy Loss, which prioritizes difficult anomalies to address class imbalance. Extensive experiments on public and industrial datasets show that ChiGAD outperforms state-of-the-art models on multiple metrics. Additionally, its homogeneous variant, ChiGNN, excels on seven GAD datasets, validating the effectiveness of Chi-Square filters. Our code is available at this https URL.
- [601] arXiv:2506.03410 (replaced) [pdf, html, other]
-
Title: An iterative tangential interpolation algorithm for model reduction of MIMO systemsComments: 14 pages, 4 figures Submitted to IEEE TAC. Revision 2Subjects: Systems and Control (eess.SY)
We consider model reduction of large-scale Multi-Input, Multi-Output (MIMO) systems using low-rank barycentric tangential interpolation on the frequency axis. Our scheme is related to the recently-developed Adaptive Antoulas--Anderson (AAA) algorithm, which is an iterative algorithm that uses concepts related to the Loewner framework.
One feature of our algorithm is optimization of the interpolation weight matrices to minimize a proxy for the \(L_2\) system error. We use the freedom in interpolation point selection to provide multiple minimum-error-based algorithms with trade-offs between computational complexity and approximation performance. We prove that a weighted \(L_2\) norm of the proxy error system is non-increasing as interpolation points are added, and that the reduced-order system is stable whenever the original system is. Computational results are provided demonstrating the closeness of the proxy system and true system \(L_2\) errors. Comparisons with prior work demonstrate performance on par with standard model reduction methods. - [602] arXiv:2506.10903 (replaced) [pdf, html, other]
-
Title: Beyond Gold Standards: Epistemic Ensemble of LLM Judges for Formal Mathematical ReasoningComments: Accepted to EACL 2027Subjects: Computation and Language (cs.CL)
Statement autoformalization plays a crucial role in formal mathematical reasoning by enabling the automatic translation of natural language statements into formal languages. While recent advances using large language models (LLMs) have shown promising capabilities in autoformalization, methods for automatically evaluating autoformalization remain underexplored. LLM-as-a-judge presents a promising approach for automating such evaluation, however, existing methods typically employ coarse-grained and generic evaluation criteria, which limit their effectiveness for advanced formal mathematical reasoning, where quality hinges on nuanced, multi-granular dimensions. In this work, we take a step toward addressing this gap by introducing a systematic, automatic method to evaluate autoformalization tasks. The proposed method is based on an epistemically and formally grounded ensemble (EFG) of LLM judges, defined on criteria encompassing logical preservation (LP), mathematical consistency (MC), formal quality (FQ), and formal validity (FV), resulting in a transparent assessment that accounts for different contributing factors. We validate the proposed framework to serve as a proxy for autoformalization assessment within the domain of formal mathematics. Overall, our experiments demonstrate that the EFG ensemble of LLM judges is a more suitable emerging proxy for evaluation than a coarse-grained model. These findings suggest that LLM judges, especially when guided by a well-defined set of atomic properties, could offer a scalable, interpretable, and reliable support for evaluating formal mathematical reasoning.
- [603] arXiv:2506.11628 (replaced) [pdf, html, other]
-
Title: Two Tiling is UndecidableComments: 38 pages, 48 figuresSubjects: Computational Geometry (cs.CG); Combinatorics (math.CO); Metric Geometry (math.MG)
We show that the following problem is undecidable: given two polygonal prototiles, determine whether the plane can be tiled with rotated and translated copies of them. This improves a result of Demaine and Langerman [SoCG 2025], who showed undecidability for three tiles.
Along the way, we show that tiling with one prototile is undecidable if there can be edge-to-edge matching rules. This is the first result to show undecidability for monotiling with only local matching constraints. - [604] arXiv:2507.09134 (replaced) [pdf, html, other]
-
Title: Integrating Planning and Predictive Control Using the Path Feasibility GovernorComments: 16 pages, 12 figures, accepted as a Full Paper in IEEE Transactions on Automatic ControlSubjects: Systems and Control (eess.SY)
The motion planning problem of generating dynamically feasible, collision-free trajectories in non-convex environments is a fundamental challenge for autonomous systems. Decomposing the problem into path planning and path tracking improves tractability, but integrating these components in a theoretically sound and computationally efficient manner is challenging. We propose the Path Feasibility Governor (PathFG), a framework for integrating path planners with nonlinear Model Predictive Control (MPC). The PathFG manipulates the reference passed to the MPC controller, guiding it along a path while ensuring constraint satisfaction, stability, and recursive feasibility. The PathFG is modular, compatible with replanning, and improves computational efficiency and reliability by reducing the need for long prediction horizons. We prove safety and asymptotic stability with a significantly expanded region of attraction, and validate its real-time performance through simulated case studies of unicycle and quadrotor navigation in cluttered environments.
\c{opyright}2026 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. - [605] arXiv:2507.10087 (replaced) [pdf, html, other]
-
Title: Foundation Model Driven Robotics: A Comprehensive ReviewSubjects: Robotics (cs.RO)
The rapid emergence of foundation models, particularly large language models and vision-language models, has introduced a transformative paradigm in robotics. The models offer powerful capabilities in semantic understanding, high-level reasoning, and cross-modal generalization, enabling significant advances in perception, planning, control, and human-robot interaction. This critical review provides a structured synthesis of recent developments and organizes applications into four directions: simulation-driven robotics, open-world operation, sim-to-real transfer, and adaptable robotic systems. Unlike existing surveys that emphasize isolated capabilities, this work highlights integrated, system-level strategies and evaluates their practical feasibility in real-world environments. Key enabling trends such as procedural scene generation, policy generalization, and multimodal reasoning are discussed alongside core bottlenecks, including limited embodiment, lack of multimodal data, safety risks, and computational constraints. Through this lens, this paper identifies both the architectural strengths and critical limitations of foundation model-based robotics, highlighting open challenges in real-time operation, grounding, resilience, and trust. The review concludes with a roadmap of proposed future research directions for bridging semantic reasoning and physical intelligence through more robust, interpretable, and embodied models.
- [606] arXiv:2507.14632 (replaced) [pdf, html, other]
-
Title: BusterX++: Towards Unified Cross-Modal AI-Generated Content Detection and Explanation with MLLMSubjects: Computer Vision and Pattern Recognition (cs.CV)
The rapid advancement of generative AI has substantially improved image and video synthesis, amplifying the risk of multimodal visual misinformation. Recent MLLMs have shown promise for transparent AI-generated content detection through reasoning and explanation, yet existing approaches largely treat image and video forensics as isolated tasks, leaving cross-modal synergies underexplored. To address this, we present \textbf{BusterX++}, a unified MLLM for joint image and video detection with interpretable reasoning. We also introduce \textbf{GenBuster-Bench++}, a curated benchmark containing balanced image and video samples spanning recent generation models and diverse real-world scenarios, with synthetic images and videos screened using the same human difficulty criterion. Using this controlled setting, we revisit the widely adopted SFT+RL post-training paradigm. Notably, in the unified image-video setting, our results demonstrate that a single-stage pure RL strategy driven primarily by sparse outcome rewards outperforms a strong SFT+RL baseline. We observe that single-stage pure RL maintains higher policy entropy than SFT-initialized RL throughout training. This observation suggests a possible role for exploration in the performance gains observed under unified image-video training. Extensive experiments demonstrate that BusterX++ achieves state-of-the-art performance, highlighting the powerful potential of RL for unified cross-modal visual reasoning.
- [607] arXiv:2508.00910 (replaced) [pdf, html, other]
-
Title: Cyber-Zero: Training Cybersecurity Agents without RuntimeComments: Public Link: this https URLSubjects: Cryptography and Security (cs.CR); Computation and Language (cs.CL); Machine Learning (cs.LG)
Large Language Models (LLMs) have achieved remarkable success in software engineering tasks when trained with executable runtime environments, particularly in resolving GitHub issues. However, such runtime environments are often unavailable in other domains, especially cybersecurity, where challenge configurations and execution contexts are ephemeral or restricted. We present Cyber-Zero, the first runtime-free framework for synthesizing high-quality agent trajectories to train cybersecurity LLMs. Cyber-Zero leverages publicly available CTF writeups and employs persona-driven LLM simulation to reverse-engineer runtime behaviors and generate realistic, long-horizon interaction sequences without actual environments. Using trajectories synthesized by Cyber-Zero, we train LLM-based agents that achieve up to 11.4% absolute performance gains over baseline models on three prominent CTF benchmarks: InterCode-CTF, NYU CTF Bench, and Cybench. Our best model, Cyber-Zero-32B, establishes new state-of-the-art performance among open-weight models, matching the capabilities of proprietary systems like DeepSeek-V3-0324 and Claude-3.5-Sonnet while offering superior cost-effectiveness, and demonstrating that runtime-free trajectory synthesis can effectively democratize the development of state-of-the-art cybersecurity agents.
- [608] arXiv:2508.01925 (replaced) [pdf, html, other]
-
Title: Robust Surrogate Modeling for Explainable Graph Neural NetworksSubjects: Machine Learning (cs.LG)
Graph Neural Networks (GNNs) have achieved remarkable performance in a wide range of graph-related learning tasks. However, explaining their predictions remains a challenging problem, especially due to the mismatch between the graphs used during training and those encountered during explanation. Most existing methods optimize soft edge masks on weighted graphs to highlight important substructures, but these graphs differ from the unweighted graphs on which GNNs are trained. This distributional shift leads to unreliable gradients and degraded explanation quality, especially when generating small, sparse subgraphs. To address this issue, we propose explaining a robust surrogate whose explanations provide a faithful proxy for those of the original model. Specifically, we train this surrogate model through an iterative re-weighting strategy that alternates between identifying explanation subgraphs and adapting the model to the resulting weighted graph distribution, supporting stable explanation refinement while preserving the validity of the soft-mask relaxation. We evaluate our method on multiple benchmark datasets using different GNN backbones and explanation methods. Experimental results show that our method consistently improves explanation quality and can be flexibly integrated with different architectures.
- [609] arXiv:2508.02358 (replaced) [pdf, html, other]
-
Title: Fully implicit timestepping methods for the rotating shallow water equationsSubjects: Numerical Analysis (math.NA); Fluid Dynamics (physics.flu-dyn)
Fully implicit timestepping methods have had limited application in geophysical fluid dynamics due to challenges of finding suitable iterative solvers, since the coupled treatment of advection prevents the standard elimination techniques. However, overlapping Additive Schwarz methods provide a robust, scalable iterative approach for solving the monolithic coupled system for all fields and Runge-Kutta stages. In this study we investigate this approach applied to the rotating shallow water equations, facilitated by the Irksome package which provides automated code generation for implicit Runge-Kutta methods. We compare various schemes in terms of accuracy and efficiency using an implicit/explicit splitting method, namely the ARK2 scheme of Giraldo et al (2013), as a benchmark. This provides an initial look at whether implicit Runge-Kutta methods can be viable for atmosphere and ocean simulation.
- [610] arXiv:2508.18104 (replaced) [pdf, html, other]
-
Title: On the Parameterized Complexity of Grundy Domination and Zero Forcing ProblemsComments: version published in the Journal of Computer and System SciencesJournal-ref: Journal of Computer and System Sciences 164, Article 103875 (2027)Subjects: Computational Complexity (cs.CC); Discrete Mathematics (cs.DM); Data Structures and Algorithms (cs.DS); Combinatorics (math.CO)
We consider two different problem families that deal with domination in graphs. On the one hand, we focus on dominating sequences. In such a sequence, every vertex dominates some vertex of the graph that was not dominated by any earlier vertex in the sequence. The problem of finding the longest dominating sequence is known as $\mathsf{Grundy~Domination}$. Depending on whether the closed or the open neighborhoods are used for domination, there are three other versions of this problem: $\mathsf{Grundy~Total~Domination}$, $\mathsf{L\text{-}Grundy~Domination}$, and $\mathsf{Z\text{-}Grundy~Domination}$. We show that all four problem variants are $\mathsf{W[1]}$-complete when parameterized by the solution size.
On the other hand, we consider the family of zero forcing problems which form the parametric duals of the Grundy domination problems. In these problems, one looks for the smallest set of vertices initially colored blue such that certain color change rules are able to color all other vertices blue. It has been shown that the dual of $\mathsf{Z\text{-}Grundy~Domination}$, known as $\mathsf{Zero~Forcing~Set}$, is in $\mathsf{FPT}$ when parameterized by the treewidth or the solution size. We extend the treewidth result to the other three variants of zero forcing and their respective Grundy domination problems. Our algorithm also implies an $\mathsf{FPT}$ algorithm for $\mathsf{Grundy~Domination}$ when parameterized by the number of vertices that are not in the dominating sequence. In contrast, we show that $\mathsf{L\text{-}Grundy~Domination}$ is $\mathsf{W[1]}$-hard for that parameter. - [611] arXiv:2508.18370 (replaced) [pdf, html, other]
-
Title: Training Language Model Agents to Find Vulnerabilities with CTF-DojoSubjects: Software Engineering (cs.SE); Computation and Language (cs.CL); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
Large language models (LLMs) have demonstrated exceptional capabilities when trained within executable runtime environments, notably excelling at software engineering tasks through verified feedback loops. Yet, scalable and generalizable execution-grounded environments remain scarce, limiting progress in training more capable ML agents. We introduce CTF-Dojo, the first large-scale executable runtime tailored for training LLMs with verifiable feedback, featuring 658 fully functional Capture-The-Flag (CTF)-style challenges containerized in Docker with guaranteed reproducibility. To enable rapid scaling without manual intervention, we develop CTF-Forge, an automated pipeline that transforms publicly available artifacts into ready-to-use execution environments in minutes, eliminating weeks of expert configuration traditionally required. We trained LLM-based agents on just 486 high-quality, execution-verified trajectories from CTF-Dojo, achieving up to 9.9% absolute gains over strong baselines across three competitive benchmarks: InterCode-CTF, NYU CTF Bench, and Cybench. Our best-performing 32B model reaches 31.9% Pass@1, establishing a new open-weight state-of-the-art that rivals frontier models like DeepSeek-V3-0324 and Gemini-2.5-Flash. By framing CTF-style tasks as a benchmark for executable-agent learning, CTF-Dojo demonstrates that execution-grounded training signals are not only effective but pivotal in advancing high-performance ML agents without dependence on costly proprietary systems.
- [612] arXiv:2509.02792 (replaced) [pdf, html, other]
-
Title: Structured Basis Function Networks: Loss-Centric Multi-Hypothesis Ensembles with Controllable DiversityComments: 50 Pages, 16 Figures, 20 TablesSubjects: Machine Learning (cs.LG)
Ensemble learning and multi-hypothesis prediction are two complementary paradigms for improving predictive performance and robustness. This work introduces the Structured Basis Function Network (s-BFN), a loss-aware ensemble framework that unifies these paradigms through a supervised aggregation mechanism motivated by the loss-induced Bregman-centroid geometry of prediction space. The approach constructs a structured dataset of base predictions and trains a parametric combiner directly against the observed labels, with the centroidal formulation serving as a geometric reference rather than an exact centroid solver, while a diversity-control parameter regulates specialization among predictors. The proposed framework is evaluated across regression and classification tasks, including tabular data, image recognition, and large-scale temporally structured video classification. Empirical results show that s-BFN provides strong performance across the evaluated regimes and is competitive with the aggregation baselines considered in the study. The strongest gains are observed in higher-dimensional and temporally structured settings, where aggregation capacity and controlled diversity are jointly beneficial, while simpler tabular regimes exhibit earlier saturation. These results highlight a capacity-complexity-diversity trade-off rather than a strictly monotonic relationship between diversity, ensemble size, and accuracy.
- [613] arXiv:2509.22482 (replaced) [pdf, html, other]
-
Title: Bayesian Transfer Operators in Reproducing Kernel Hilbert SpacesSubjects: Machine Learning (cs.LG); Dynamical Systems (math.DS); Chaotic Dynamics (nlin.CD); Data Analysis, Statistics and Probability (physics.data-an)
Transfer operators provide a linear description of nonlinear dynamical systems by propagating observables or probability distributions. In reproducing kernel Hilbert spaces, these operators are represented through conditional mean embeddings, whose regularized estimator coincides with the posterior mean of Gaussian process regression. Building on this correspondence, we formulate Gaussian Process Dynamic Mode Decomposition, a Bayesian transfer-operator method that combines sparse variational Gaussian processes with kernel-based Koopman and Perron--Frobenius operator approximation. The resulting framework provides a principled route to uncertainty quantification, scalable sparse approximations, and data-adaptive dictionary learning for prediction and spectral analysis. Additionally, the model-selection formulation naturally gives rise to a forward--backward operator structure, which establishes a connection between Bayesian transfer-operator learning and time-lagged canonical correlation analysis.
- [614] arXiv:2509.23373 (replaced) [pdf, html, other]
-
Title: Graph Your Own PromptComments: NeurIPS 2025 (Author-corrected version: Revised by the original authors, October 2026)Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
We propose Graph Consistency Regularization (GCR), a novel framework that injects relational graph structures, derived from model predictions, into the learning process to promote class-aware, semantically meaningful feature representations. Functioning as a form of self-prompting, GCR enables the model to refine its internal structure using its own outputs. While deep networks learn rich representations, these often capture noisy inter-class similarities that contradict the model's predicted semantics. GCR addresses this issue by introducing parameter-free Graph Consistency Layers (GCLs) at arbitrary depths. Each GCL builds a batch-level feature similarity graph and aligns it with a global, class-aware masked prediction graph, derived by modulating softmax prediction similarities with intra-class indicators. This alignment enforces that feature-level relationships reflect class-consistent prediction behavior, acting as a semantic regularizer throughout the network. Unlike prior work, GCR introduces a multi-layer, cross-space graph alignment mechanism with adaptive weighting, where layer importance is learned from graph discrepancy magnitudes. This allows the model to prioritize semantically reliable layers and suppress noisy ones, enhancing feature quality without modifying the architecture or training procedure. GCR is model-agnostic, lightweight, and improves semantic structure across various networks and datasets. Experiments show that GCR promotes cleaner feature structure, stronger intra-class cohesion, and improved generalization, offering a new perspective on learning from prediction structure. [Project website](this https URL) [Code](this https URL)
- [615] arXiv:2510.01381 (replaced) [pdf, html, other]
-
Title: A Stochastic Framework for Continuous-Time State Estimation of Continuum RobotsComments: 20 pages, 15 figures. Accepted to IEEE Transactions on Robotics. Notable changes from v1 include switching to a GTSAM back-end, adjusting the spatial factor definition, and the inclusion of a sensitivity studySubjects: Robotics (cs.RO)
State estimation techniques for continuum robots (CRs) typically involve using computationally complex dynamic models, simplistic shape approximations, or are limited to quasi-static methods. These approaches can be sensitive to unmodeled disturbances acting on the robot. Inspired by a factor-graph optimization paradigm, this work introduces a continuous-time stochastic state estimation framework for continuum robots. We introduce factors based on continuous-time kinematics that are corrupted by a white noise Gaussian process (GP). By using a simple robot model paired with high-rate sensing, we show adaptability to unmodeled external forces and data dropout. The result contains an estimate of the mean and covariance for the robot's pose, velocity, and strain, each of which can be interpolated continuously in both space and time. This same interpolation scheme can be used during estimation, allowing for inclusion of measurements that are asynchronous to the estimated states. Our method's inherent sparsity leads to a linear solve complexity with respect to time and interpolation queries in constant time. We demonstrate our method on a CR with gyroscope and pose sensors, highlighting its versatility in real-world systems.
- [616] arXiv:2510.05241 (replaced) [pdf, html, other]
-
Title: Simultaneous Learning and Optimization via Misspecified Saddle Point ProblemsComments: Accepted for publication in the Journal of Computational Optimization and AlgorithmsSubjects: Machine Learning (cs.LG); Optimization and Control (math.OC)
We study a class of misspecified saddle point (SP) problems, where the optimization objective depends on an unknown parameter that must be learned concurrently from data. Unlike existing studies that assume parameters are fully known or pre-estimated, our framework integrates optimization and learning into a unified formulation, enabling a more flexible problem class. To address this setting, we propose two algorithms based on the accelerated primal-dual (APD) by Hamedani & Aybat 2021. In particular, we first analyze the naive extension of the APD method by directly substituting the evolving parameter estimates into the primal-dual updates; then, we design a new learning-aware variant of the APD method that explicitly accounts for parameter dynamics by adjusting the momentum updates. Both methods achieve a provable convergence rate of $\mathcal{O}(\log K / K)$, while the learning-aware approach attains a tighter $\mathcal{O}(1)$ constant and further benefits from an adaptive step-size selection enabled by a backtracking strategy. Furthermore, we extend the framework to problems where the learning problem admits multiple optimal solutions, showing that our modified algorithm for a structured setting achieves an $\mathcal{O}(1/\sqrt{K})$ rate. To demonstrate practical impact, we evaluate our methods on a misspecified portfolio optimization problem and show superior empirical performance compared to state-of-the-art algorithms.
- [617] arXiv:2510.08697 (replaced) [pdf, html, other]
-
Title: BigCodeArena: Unveiling More Reliable Human Preferences in Code Generation via ExecutionTerry Yue Zhuo, Xiaolong Jin, Hange Liu, Juyong Jiang, Tianyang Liu, Chen Gong, Bhupesh Bishnoi, Vaisakhi Mishra, Marek Suppa, Noah Ziems, Saiteja Utpala, Ming Xu, Guangyu Song, Kaixin Li, Yuhan Cao, Bo Liu, Zheng Liu, Sabina Abdurakhmanova, Wenhao Yu, Mengzhao Jia, Jihan Yao, Kenneth Hamilton, Kumar Shridhar, Minh Chien Vu, Dingmin Wang, Jiawei Liu, Zijian Wang, Qian Liu, Binyuan Hui, Meg Risdal, Ahsen Khaliq, Atin Sood, Zhenchang Xing, Wasi Uddin Ahmad, John Grundy, David Lo, Banghua Zhu, Xiaoning Du, Torsten Scholak, Leandro von WerraComments: Built with love by the BigCode community :)Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Crowdsourced model evaluation platforms, such as Chatbot Arena, enable real-time evaluation from human perspectives to assess the quality of model responses. In the coding domain, manually examining the quality of LLM-generated content is extremely challenging, as it requires understanding long chunks of raw code and deliberately simulating code execution. To this end, we introduce BigCodeArena, an open human evaluation platform for code generation backed by a comprehensive and on-the-fly execution environment. Built on top of Chatbot Arena, BigCodeArena enables the execution of LLM-generated code and allows humans to interact with the execution process and outcomes. We collected over 14,000 raw code-centric conversation sessions across 10 widely used LLMs, spanning 10 languages and 8 types of execution environments. Among these conversations, we identified more than 4,700 multi-turn samples with pairwise human preferences. Further analysis uncovers underexplored preferences of LLMs in fine-grained domains characterized by tasks, languages, and frameworks. To systematically examine code understanding and generation capabilities of frontier LLMs, we curated two benchmarks based on the collected data, namely BigCodeReward and AutoCodeArena. For BigCodeReward, we post-processed the 4,700 conversations and evaluated the consistency between reward models and human preferences. The evaluation shows that most LLMs have superior performance in judging coding preferences when the execution results are available. Inspired by these findings, we propose AutoCodeArena, an automatic Elo rating benchmark designed to assess the coding quality of LLMs without human involvement. We find that proprietary LLMs like GPT-5, Claude-Sonnet-4, and Claude-Opus-4 still lead in code generation performance among recent emerging models.
- [618] arXiv:2510.10140 (replaced) [pdf, html, other]
-
Title: Adversarial Attacks on Downstream Weather Forecasting Models: Application to Tropical Cyclone Trajectory PredictionComments: Compared with the previous version, we (a) extend the downstream attack to an end-to-end attack; (b) further analyze physical feasibility and attack difficulty; and (c) develop a defense mechanismSubjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
Deep learning-based weather forecasting (DLWF) models leverage past weather observations to generate future forecasts, supporting a wide range of downstream applications, including tropical cyclone (TC) prediction. In this paper, we investigate the vulnerability of these models to adversarial attacks, where subtle perturbations to upstream forecasts can alter downstream TC trajectory predictions. Crafting perturbed inputs to steer downstream outputs toward attacker-specified trajectories is a challenging problem for several reasons. First, conventional TC detection systems are opaque, non-differentiable black boxes, making standard gradient-based attacks infeasible. Second, the extreme rarity of TC events leads to a severe class imbalance problem, making it difficult to develop attack methods that produce realistic-looking cyclone paths aligned with target trajectories. To overcome these limitations, we propose Cyc-Attack, a novel method for perturbing the forecasts of DLWF models to generate adversarial trajectories. The proposed method uses a differentiable surrogate model to approximate the TC detector's output, enabling the application of gradient-based attacks. Cyc-Attack also employs a skewness-aware loss function with a kernel dilation strategy to address the severe class imbalance problem. Distance-based gradient weighting and regularization strategies are also introduced to constrain perturbations to eliminate unrealistic-looking TC paths. Our experiments show that Cyc-Attack achieves a higher true positive rate in matching the target trajectories, along with lower false alarm rates and stealthier perturbations than conventional attack methods. We further demonstrate that the adversarial upstream forecasts generated by Cyc-Attack can be leveraged as initial guidance to train an end-to-end model for learning adversarial input samples that would induce target trajectories.
- [619] arXiv:2510.12200 (replaced) [pdf, html, other]
-
Title: HackWorld: Evaluating Computer-Use Agents on Exploiting Web Application VulnerabilitiesXiaoxue Ren, Penghao Jiang, Kaixin Li, Zhiyong Huang, Xiaoning Du, Jiaojiao Jiang, Zhenchang Xing, Jiamou Sun, Terry Yue ZhuoSubjects: Cryptography and Security (cs.CR); Computation and Language (cs.CL)
Web applications are prime targets for cyberattacks as gateways to critical services and sensitive data. Traditional penetration testing is costly and expertise-intensive, making it difficult to scale with the growing web ecosystem. While language model agents show promise in cybersecurity, modern web applications demand visual understanding, dynamic content handling, and multi-step interactions that only computer-use agents (CUAs) can perform. Yet, their ability to discover and exploit vulnerabilities through graphical interfaces remains largely unexplored. We present HackWorld, the first framework for systematically evaluating CUAs' capabilities to exploit web application vulnerabilities via visual interaction. Unlike sanitized benchmarks, HackWorld includes 36 real-world applications across 11 frameworks and 7 languages, featuring realistic flaws such as injection vulnerabilities, authentication bypasses, and unsafe input handling. Using a Capture-the-Flag (CTF) setup, it tests CUAs' capacity to identify and exploit these weaknesses while navigating complex web interfaces. Evaluation of state-of-the-art CUAs reveals concerning trends: exploitation rates below 12% and low cybersecurity awareness. CUAs often fail at multi-step attack planning and misuse security tools. These results expose the current limitations of CUAs in web security contexts and highlight opportunities for developing more security-aware agents capable of effective vulnerability detection and exploitation.
- [620] arXiv:2510.14893 (replaced) [pdf, html, other]
-
Title: STITCHER: Constrained Trajectory Planning in Complex Environments with Real-Time Motion Primitive SearchSubjects: Robotics (cs.RO)
Autonomous high-speed navigation through large, complex environments requires real-time generation of agile trajectories that are dynamically feasible, collision-free, and satisfy state or actuator constraints. Modern trajectory planning techniques primarily use numerical optimization, as they enable the systematic computation of high-quality, expressive trajectories that satisfy various constraints. However, stringent requirements on computation time and the risk of numerical instability can limit the use of optimization-based planners in safety-critical scenarios. This work presents an optimization-free planning framework called STITCHER that stitches short trajectory segments together with graph search to compute long-range, expressive, and near-optimal trajectories in real-time. STITCHER outperforms modern optimization-based planners through our innovative planning architecture and several algorithmic developments that make real-time planning possible. Extensive simulation testing is performed to analyze the algorithmic components that make up STITCHER, along with a thorough comparison with three state-of-the-art optimization planners. Simulation tests show that safe trajectories can be created within a few milliseconds for paths that span the entirety of two 50 m x 50 m environments. Hardware tests with a custom quadrotor verify that STITCHER can produce trackable paths in real-time while respecting nonconvex constraints, such as limits on tilt angle and motor forces, with flight speeds up to 63 km/h.
- [621] arXiv:2510.17039 (replaced) [pdf, other]
-
Title: Click, Predict, Trust: Clinician-Guided and Clinician-Evaluated AI Segmentation for Lung Cancer CT-Based Prognosis within the Knowledge-to-Action FrameworkMohammad R. Salmanpour, Sonya Falahati, Amir Hossein Pouria, Amin Mousavi, Somayeh Sadat Mehrnia, Morteza Alizadeh, Arman Gorji, Zeinab Farsangi, Alireza Safarian, Mehdi Maghsudi, Carlos Uribe, Arman Rahmim, Ren YuanComments: 13 pages, 2 figures, and 2 tablesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Lung cancer remains the leading cause of cancer mortality, with CT imaging central to screening, prognosis, and treatment. Manual segmentation is variable and time-intensive, while deep learning (DL) offers automation but faces barriers to clinical adoption. Guided by the Knowledge-to-Action framework, this study develops a clinician-in-the-loop DL pipeline to enhance reproducibility, prognostic accuracy, and clinical trust. Multi-center CT data from 999 patients across 12 public datasets were analyzed using five DL models (3D Attention U-Net, ResUNet, VNet, ReconNet, SAM-Med3D), benchmarked against expert contours on whole and click-point cropped images. Segmentation reproducibility was assessed using 497 PySERA-extracted radiomic features via Spearman correlation, ICC, Wilcoxon tests, and MANOVA, while prognostic modeling compared supervised (SL) and semi-supervised learning (SSL) across 38 dimensionality reduction strategies and 24 classifiers. Six physicians qualitatively evaluated masks across seven domains, including clinical meaningfulness, boundary quality, prognostic value, trust, and workflow integration. VNet achieved the best performance (Dice = 0.83, IoU = 0.71), radiomic stability (mean correlation = 0.76, ICC = 0.65), and predictive accuracy under SSL (accuracy = 0.88, F1 = 0.83). SSL consistently outperformed SL across models. Radiologists favored VNet for peritumoral representation and smoother boundaries, preferring AI-generated initial masks for refinement rather than replacement. These results demonstrate that integrating VNet with SSL yields accurate, reproducible, and clinically trusted CT-based lung cancer prognosis, highlighting a feasible path toward physician-centered AI translation.
- [622] arXiv:2510.18691 (replaced) [pdf, html, other]
-
Title: Investigating LLM Capabilities on Long Context Comprehension for Clinical Question AnsweringSubjects: Computation and Language (cs.CL)
Despite the impressive reported performance of Large Language Models (LLMs) on medical Question-Answering (QA), evaluation still overrelies on multiple-choice and extractive settings primarily assessing textbook knowledge, with limited work on open-ended QA over patient records. We present a comprehensive study investigating LLM long-context (LC) comprehension on longitudinal, patient-centric Electronic Health Records (EHRs) across multiple datasets, models and metrics, covering extractive, multiple-choice, and open-ended generative QA under different context formulation and Retrieval Augmented Generation (RAG) settings. We show that medically fine-tuned models offer no measurable benefit over base counterparts, smaller models exhibit stronger memorization, and longitudinal supplementary notes aid generative QA. Reasoning models gain on complex multi-note QA under LLM Correctness but not stricter reference-aligned metrics, while hybrid RAG with reranking outperforms Full-Context (FC) prompting. A question-category analysis and multi-faceted metric evaluation further reveal where RAG and FC excel and struggle, with RAG falling behind in longitudinal reasoning, thus uncovering systematic failures pointing to the need for reasoning-appropriate metrics.
- [623] arXiv:2510.19244 (replaced) [pdf, html, other]
-
Title: Interpret Policies in Deep Reinforcement Learning using SILVER with RL-Guided Labeling: A Model-level Approach to High-dimensional and Multi-action EnvironmentsSubjects: Machine Learning (cs.LG)
Deep reinforcement learning (RL) achieves remarkable performance but lacks interpretability, limiting trust in policy behavior. The existing SILVER framework (Li, Siddique, and Cao 2025) explains RL policy via Shapley-based regression but remains restricted to low-dimensional, binary-action domains. We propose SILVER with RL-guided labeling, an enhanced variant that extends SILVER to multi-action and high-dimensional environments by incorporating the RL policy's own action outputs into the boundary points identification. Our method first extracts compact feature representations from image observations, performs SHAP-based feature attribution, and then employs RL-guided labeling to generate behaviorally consistent boundary datasets. Surrogate models, such as decision trees and regression-based functions, are subsequently trained to interpret RL policy's decision structure. We evaluate the proposed framework on two Atari environments using three deep RL algorithms and conduct human-subject study to assess the clarity and trustworthiness of the derived interpretable policy. Results show that our approach maintains competitive task performance while substantially improving transparency and human understanding of agent behavior. This work advances explainable RL by transforming SILVER into a scalable and behavior-aware framework for interpreting deep RL agents in high-dimensional, multi-action settings.
- [624] arXiv:2510.20279 (replaced) [pdf, html, other]
-
Title: CS-54k: A Paper-Grounded Corpus and Benchmark for AI Research QAPenghao Wang, Yuhao Zhou, Mengxuan Wu, Ziheng Qin, Bangyuan Zhu, Shengbin Huang, Xuanlei Zhao, Panpan Zhang, Xiaojiang Peng, Yuzhang Shang, Jianfei Yang, Zheng Zhu, Tianlong Chen, Zhangyang Wang, Kai WangSubjects: Machine Learning (cs.LG)
Large language models (LLMs) are increasingly being used to support computer science research. To contribute reliably, these systems must be able to work with information distributed across research papers and across different aspects of the research process. However, large-scale resources that systematically organize such information into structured supervision for both model evaluation and post-training remain limited. To address this gap, we propose a systematic paper-to-QA construction pipeline that organizes information from AI research papers into structured QA pairs supported by source evidence. Using this pipeline, we construct CS-54k, a 54k-pair corpus derived from 14,474 openly licensed AI research papers. The corpus covers eight recurring paper components, including research domain, prior work, existing challenges, motivation, findings and assumptions, methods, experimental settings, and experimental results. We further derive two complementary subsets: CS-4k, a benchmark for evaluating LLMs on paper-derived research QA, and CS-50k, a companion corpus for post-training. Extensive evaluations on CS-4k reveal substantial performance differences across existing LLMs and paper components, with experimental settings and results remaining particularly challenging. Moreover, post-training on CS-50k substantially improves Qwen2.5-7B-Instruct on CS-4k, demonstrating the practical utility of structured, paper-derived supervision. Together, CS-4k and CS-50k provide a unified and scalable resource for benchmarking and improving LLMs on research paper question answering.
- [625] arXiv:2510.22953 (replaced) [pdf, html, other]
-
Title: Manifold Approximation leads to Robust Kernel AlignmentComments: 9 pages, 5 figures + supplementaryJournal-ref: Transactions on Machine Learning Research, 2026Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
Centered kernel alignment (CKA) is a popular metric for comparing representations, determining equivalence of networks, and conducting neuroscience research. However, CKA does not account for the underlying manifold and relies on numerous heuristics that cause it to behave differently at different scales of data. In this work, we propose Manifold-approximated Kernel Alignment (MKA), which incorporates manifold geometry (i.e., local neighborhood) into the alignment task. We derive a theoretical framework for MKA. Our construction removes representation-dependent degree weighting, yields a closed-form alignment score, and gives each row a probabilistic interpretation as a local Markov transition distribution. We further establish kernel- and score-level stability with respect to the neighborhood size. We perform empirical evaluations on synthetic datasets and real-world examples to characterize and compare MKA to its contemporaries. In controlled synthetic experiments, MKA is consistently less sensitive to hyperparameters and captures the selected topology-preserving relationships. On the large-scale benchmark, MKA achieves its strongest aggregate performance in vision, is competitive on graphs, and is weaker than CKA in several NLP evaluations. Our findings suggest that manifold-aware kernel alignment provides a robust foundation for measuring representations and is potentially useful when local neighborhood geometry is important, with possible applications in representation learning.
- [626] arXiv:2511.11131 (replaced) [pdf, html, other]
-
Title: Expert-Guided Online Reinforcement Learning for Stable and Sample-Efficient Control of Dynamical SystemsComments: Submitted to L4DCSubjects: Systems and Control (eess.SY)
Online reinforcement learning (RL) often requires extensive interactions with the environment, which makes the training costly. Additionally, exploration can destabilise the dynamical system, which may be unsafe. On the other hand, offline learning from expert demonstrations avoids unsafe exploration but is limited by the quality and coverage of the offline dataset. In this paper, we propose a model-free RL approach that combines expert demonstrations with limited online interactions. Specifically, we optimise a policy objective consisting of an online performance term and a behavioural cloning regulariser computed from offline expert demonstrations. The regulariser encourages the learned policy to stay close to expert behaviour, reducing the need to explore unfamiliar actions. Expert demonstrations guide policy learning and reduce the need for many online interactions and exploration. For stochastic linear time-invariant (LTI) systems, we establish conditions for linear convergence of exact policy gradient updates and for all intermediate policies to remain stabilising. Although our theoretical analysis focuses on linear systems, we also evaluate the proposed approach on nonlinear systems to demonstrate its effectiveness beyond the linear setting. \rev{Numerical results show that the proposed method reaches comparable performance in fewer gradient steps and with less number of samples than a purely online RL baseline.
- [627] arXiv:2511.12069 (replaced) [pdf, other]
-
Title: A GNN-Based Approach for Code Smell Detection and Refactoring Opportunity IdentificationComments: We would like to withdraw our manuscript to make further revisions and improvements to its content and experimental analysis. We apologize for any inconvenience and appreciate your understandingSubjects: Software Engineering (cs.SE); Methodology (stat.ME)
Code smells indicate latent design or implementation problems that may negatively affect software maintainability. Over the past decades, numerous approaches have been proposed for code smell refactoring, including those based on software metrics, predefined rules, machine learning/deep learning (ML/DL), and, more recently, large language models (LLMs). However, traditional metric- and rule-based approaches rely heavily on manually defined criteria, while ML/DL approaches are often constrained by the availability of training data and the design of model architectures. Although LLMs provide advanced code understanding capabilities, they typically incur substantial computational and time costs, particularly when applied to large-scale software systems. In this study, we propose a graph-based deep learning approach for code smell refactoring. Specifically, we construct code graphs at both the class and method levels and employ graph- and node-level classification to detect code smells and identify corresponding refactoring opportunities for three representative code smells: Long Method, Large Class, and Feature Envy. The proposed approach is trained and evaluated using SACS, a large-scale open-source dataset constructed through a semi-automated process. In the experiments, we implement the proposed approach using three representative GNN architectures: GCN, GraphSAGE, and GAT. We further compare its performance with traditional refactoring approaches, an existing deep learning-based approach, and four representative lightweight LLMs. Experimental results show that the proposed approach outperforms the compared approaches across all evaluated code smell refactoring tasks, demonstrating that it provides both superior performance and greater efficiency.
- [628] arXiv:2511.21542 (replaced) [pdf, html, other]
-
Title: E0: Expressive Fine-Grained Discrete Action Prediction for Vision-Language-Action Models via Tweedie Discrete DiffusionZhihao Zhan, Jiaying Zhou, Likui Zhang, Qinhan Lyu, Hao Liu, Weizheng Li, Jusheng Zhang, Ziliang Chen, Tianshui Chen, Ruifeng Zhai, Keze Wang, Liang Lin, Guangrun WangComments: Accepted to NeurIPS 26Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Vision-Language-Action (VLA) models offer a unified framework for robotic manipulation by integrating visual perception, language understanding, and control generation. However, existing VLA systems still struggle to generalize across diverse tasks, scenes, and camera viewpoints, and often produce coarse or unstable actions. We argue that these limitations are closely tied to the structural properties of actions in VLA settings, including the inherent multi-peaked nature of action distributions, the token-based symbolic reasoning of pretrained VLM/VLA backbones, and the effective finite resolution imposed by real-world robotic control. Motivated by these properties, we introduce E0, a tweedie discrete diffusion framework that formulates action generation as iterative denoising over quantized action tokens. By operating in a discrete action space with a principled diffusion process, E0 naturally aligns with token-based reasoning, supports fine-grained yet executable action control, and avoids the distributional mismatch of masking-based discrete diffusion. Experiments on LIBERO, VLABench, ManiSkill, and a real-world Franka arm demonstrate that E0 achieves state-of-the-art performance across diverse environments, outperforming strong baselines by 8.1% on average.
- [629] arXiv:2511.23385 (replaced) [pdf, html, other]
-
Title: Strong nonlinear detectability and moving horizon estimation for nonlinear systems with unknown inputsComments: 19 pages,3 figuresSubjects: Systems and Control (eess.SY)
This paper investigates state estimation for general nonlinear discrete-time systems subject to measurement noise and possibly unbounded unknown inputs. To approach this problem, we first propose the concept of strong nonlinear detectability, which proves to be a nonlinear counterpart to strong* detectability. Based on the proposed detectability notion, an unknown input state estimator (UISE), which yields bounded estimation error even for unbounded unknown inputs, is designed via a moving horizon estimation strategy using past and current measurements. Next, we tighten this detectability notion to develop a two-stage MHE-based UISE that is computationally more efficient. In a simulation example, both variants of MHE-based UISEs are compared with a conventional MHE to illustrate the merits of the proposed methods.
- [630] arXiv:2512.03733 (replaced) [pdf, html, other]
-
Title: A Superfast Direct Solver for Type-III Inverse Nonuniform Discrete Fourier TransformSubjects: Numerical Analysis (math.NA)
Various fields in scientific computing require the use of the nonuniform discrete Fourier transform (NUDFT) and its inversion. In [H. Wilber, E. N. Epperly, and A. H. Barnett, SIAM J. Sci. Comput., 47 (2025), pp. A1702-A1732], the authors proposed a direct inversion method for the type-II NUDFT, which achieves quasi-linear complexity by approximating the type-II NUDFT matrix as a product of a hierarchical semiseparable (HSS) matrix and a DFT matrix. In this article, we propose a novel superfast direct inversion method for type-III NUDFT. The proposed method approximates the type-III NUDFT matrix as a product of a type-II NUDFT matrix and an HSS matrix. This decomposition enables both the forward application and the backward inversion to be accomplished with quasi-linear complexity. The fast inversion can serve as a high-accuracy direct solver or as an efficient preconditioner. Additionally, we provide an error bound for the approximation when the sample points are independent and identically distributed uniform random variables and the frequencies are perturbed continuous integers. Numerical results are presented to verify the relevant theoretical properties and demonstrate the efficiency of the proposed methods.
- [631] arXiv:2512.04821 (replaced) [pdf, html, other]
-
Title: LatentFM: A Latent Flow Matching Approach for Generative Medical Image SegmentationComments: Best Paper Award, FISU Joint Conference on Artificial Intelligence 2026 (FJCAI 2026)Journal-ref: Proceedings of the FISU Joint Conference on Artificial Intelligence 2026 (FJCAI 2026), pp. 288-294, ISBN: 978-604-45-2586-0Subjects: Computer Vision and Pattern Recognition (cs.CV)
Generative models have achieved remarkable progress with the emergence of flow matching (FM). It has demonstrated strong generative capabilities and attracted significant attention as a simulation-free flow-based framework capable of learning exact data densities. Motivated by these advances, we propose LatentFM, a flow-based model operating in the latent space for medical image segmentation. To model the data distribution, we first design two variational autoencoders (VAEs) to encode both medical images and their corresponding masks into a lower-dimensional latent space. We then estimate a conditional velocity field that guides the flow based on the input image. By sampling multiple latent representations, our method synthesizes diverse segmentation outputs whose pixel-wise variance reliably captures the underlying data distribution, enabling both highly accurate and uncertainty-aware predictions. Furthermore, we generate confidence maps that quantify the model certainty, providing clinicians with richer information for deeper analysis. We conduct experiments on two datasets, ISIC-2018 and CVC-Clinic, and compare our method with several prior baselines, including both deterministic and generative approach models. Through comprehensive evaluations, both qualitative and quantitative results show that our approach achieves superior segmentation accuracy while remaining highly efficient in the latent space.
- [632] arXiv:2512.04954 (replaced) [pdf, html, other]
-
Title: Amortized Inference of Multi-Modal Posteriors using Likelihood-Weighted Normalizing FlowsComments: 34 pages, 12 figures, 8 tablesSubjects: Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); High Energy Physics - Phenomenology (hep-ph); Computational Physics (physics.comp-ph); Data Analysis, Statistics and Probability (physics.data-an)
We present a technique for posterior estimation using Normalizing Flows trained with likelihood-weighted importance sampling, the special case of neural importance sampling in which the proposal is the prior. This approach allows for the efficient inference of theoretical parameters in high-dimensional inverse problems without the need for posterior training samples. We implement the method on multi-modal benchmark tasks in 2D and 3D to check for the efficacy. A critical observation of our study is the impact of the base distribution on the modelled posteriors. We find that standard unimodal base distributions leave spurious probability \textit{bridges} between well-separated modes. We demonstrate that initializing the flow with a Gaussian Mixture Model that matches the number of target modes significantly improves reconstruction fidelity, as measured by some distance and divergence metrics. Finally, we apply this method to a curated problem in heavy flavour physics --- the extraction of the Wolfenstein parameters from the CP asymmetry in $B^0\to J/\psi\,K^0$; it is multimodal, non-Gaussian, and asymmetric in its mode weights --- and compare the results against a well-converged Markov Chain Monte Carlo reference and an exact reference computed by quadrature, using different metrics. Used as a proposal for importance sampling, the trained flow removes its remaining bias and gives the evidence.
- [633] arXiv:2512.05844 (replaced) [pdf, html, other]
-
Title: NEAT: Neighborhood-Guided, Efficient, Autoregressive Set Transformer for 3D Molecular GenerationComments: NeurIPS 2026Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Transformer-based autoregressive models offer an efficient alternative to diffusion- and flow-based approaches for generating 3D molecules. One challenge remains: standard transformer architectures require a sequential ordering of tokens, which is not inherently defined for the atoms in a molecule. Prior works have addressed this by using canonical atom orderings. However, these approaches are not permutation invariant w.r.t. atoms and bias next-token prediction towards ordering conventions. We overcome this limitation by introducing a novel neighborhood-guided training strategy. Our model, NEAT (Neighborhood-Guided, Efficient, Autoregressive Set Transformer) treats molecular graphs as sets of atoms and learns an order-agnostic distribution over admissible tokens at the graph boundary, thereby ensuring atom-level permutation invariance. NEAT achieves state-of-the-art generation quality on the QM9 and GEOM-Drugs datasets, enables seamless completion of molecular prefixes, and offers a significant speed advantage over existing baselines.
- [634] arXiv:2512.08867 (replaced) [pdf, html, other]
-
Title: From Conversation to Evaluation: Benchmarking LLMs on Development Knowledge via SimpleDevQAJing Zhang, Lianghong Guo, Yanlin Wang, Terry Yue Zhuo, Yong Wang, Mingwei Liu, Jiachi Chen, Ensheng Shi, Yuchi Ma, Hongyu Zhang, Zibin ZhengSubjects: Software Engineering (cs.SE)
The Development Knowledge Question Answering (Dev Knowledge QA) task aims to provide accurate natural language answers to knowledge-seeking questions during software development. To investigate the importance of Dev Knowledge QA in AI-assisted software development and the extent to which it has been explored, we conduct a preliminary analysis of real user-LLM dialogues from WildChat. Our findings indicate that Dev Knowledge QA plays a significant role in real-world software development scenarios, and these raw dialogues cannot be directly used to construct a Dev Knowledge QA benchmark. Existing Dev Knowledge QA benchmarks are limited in development knowledge scope and often not built from real user queries. To bridge this gap, we design a three-phase pipeline that transforms real-world dialogue into simple development knowledge-seeking QA pairs. Through this pipeline, we introduce SimpleDevQA, a multilingual Dev Knowledge QA benchmark inspired by real user dialogues. This dataset covers three languages (English, Chinese, and Russian), and focuses on questions with unique, short, and verifiable answers, making evaluation more accurate and simple. Extensive experiments with 21 mainstream LLMs show that closed-source models generally perform best on SimpleDevQA. We also find that RAG-based knowledge injection improves accuracy, and that Dev Knowledge QA performance correlates with both model confidence and code-generation capability. To facilitate the replication study, we have released our data and code at: this https URL.
- [635] arXiv:2512.20777 (replaced) [pdf, html, other]
-
Title: Improving Matrix Exponential for Generative AI Flows: A Taylor-Based Approach Beyond Paterson--StockmeyerComments: 52 pages, 4 figures. Revised version: added new theoretical results, sharper error analysis, additional numerical experiments, benchmarking on modern GPU hardware, comparisons with standard library implementations, and validation of generative model qualitySubjects: Machine Learning (cs.LG); Numerical Analysis (math.NA)
The matrix exponential is a fundamental operator in scientific computing and system simulation, with applications ranging from control theory and quantum mechanics to modern generative machine learning. While Padé approximants combined with scaling and squaring have long served as the standard, recent Taylor-based methods, which utilize polynomial evaluation schemes that surpass the classical Paterson--Stockmeyer technique, offer superior accuracy and reduced computational complexity. This paper presents an optimized Taylor-based algorithm for the matrix exponential, specifically designed for the high-throughput requirements of generative AI flows. We provide a rigorous error analysis and develop a dynamic selection strategy for the Taylor order and scaling factor to minimize computational effort under a prescribed error tolerance. Extensive numerical experiments demonstrate that our approach provides significant acceleration and maintains high numerical stability compared to existing state-of-the-art implementations. These results establish the proposed method as a highly efficient tool for large-scale generative modeling.
- [636] arXiv:2601.01547 (replaced) [pdf, html, other]
-
Title: Vision-language models lag human performance on physical dynamics and intent reasoningComments: 29 pages, 7 figures. Supplementary Information and Source Data are included as ancillary filesSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Spatial intelligence is central to embodied cognition, yet contemporary AI systems still struggle to reason about physical interactions in open-world human environments. Despite strong performance on controlled benchmarks, vision-language models often fail to jointly model physical dynamics, reference frames, and the latent human intentions that drive spatial change. We introduce Teleo-Spatial Intelligence (TSI), which captures an important aspect of spatial reasoning: integrating physical dynamics and action goals to reason about actions and their spatial consequences. To evaluate TSI, we present EscherVerse, a large-scale open-world resource built from 11,328 real-world videos, including an 8,000-example benchmark and a 35,963-example instruction-tuning set. Across 27 state-of-the-art vision-language models and an independent analysis of first-pass human responses from 11 annotators, we identify a persistent teleo-spatial reasoning gap: the strongest proprietary model achieves 57.26% overall accuracy, far below first-pass human performance, which ranges from 84.81% to 95.14% with a mean of 90.62%. Fine-tuning on real-world, intent-aware data narrows this gap for open-weight models, but does not close it. EscherVerse provides a diagnostic testbed for purpose-aware spatial reasoning and highlights a critical gap between pattern recognition and human-level understanding in embodied AI.
- [637] arXiv:2601.08113 (replaced) [pdf, html, other]
-
Title: Coordinated Cooling and Compute Management for AI DatacentersComments: Accepted at IEEE Transactions on Cloud ComputingSubjects: Systems and Control (eess.SY); Distributed, Parallel, and Cluster Computing (cs.DC)
The AI datacenters are currently being deployed on a large scale to support the training and deployment of power-intensive large-language models (LLMs). Extensive amount of computation and cooling required in datacenters increase concerns about the energy use and carbon emissions of AI datacenters. Although current state-of-the-art has examined the energy efficiency of LLM inference, most prior research focused on optimizing compute-side scheduling without considering thermal objectives or constraints. Since GPU-intensive inference generates substantial heat that can degrade datacenter performance, ignoring thermal effects can increase total energy consumption and reduce the efficiency of LLM serving. To fill this gap, we profile the characteristics of GPU servers under varying cooling and AI jobs, and develop a joint cooling and computing modeling approach for AI datacenters. Built upon such workload and thermal dynamics models, a novel hierarchical control framework is proposed to co-optimize computing and thermal management by identifying the optimal GPU parallelism, frequency (DVFS), and cooling control knobs. Using real Azure inference traces and detailed GPU profiling, our model balances serving latency and thermal constraints in AI datacenters while significantly improving AI datacenters' energy efficiency.
- [638] arXiv:2601.08732 (replaced) [pdf, html, other]
-
Title: ISLA: Ensemble learning and domain adaptation for generalizable acute ischemic stroke lesion segmentation in diffusion MRI with clinical validationVincent Roca, Hakim Daouairi, Sevak Yepremian, Hilde Henon, Nicolas Martinage, Laurent Puy, Grégory Kuchcinski, Martin Bretzner, Renaud LopesComments: Substantially revised manuscript, including improvements to the methodology, additional experiments, and an expanded discussionSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Accurate segmentation of acute ischemic stroke (AIS) lesions on MRI is essential for clinical decision-making. Deep learning methods offer a promising approach to automating this task. However, they often generalize poorly to unseen clinical sites, and few AIS lesion segmentation methods have been externally validated. In this study, we propose $\mathit{ISLA}$ (Ischemic Stroke Lesion Analyzer), a framework for AIS lesion segmentation built on the design principles of $\mathit{nnU\text{-}Net}$. $\mathit{ISLA}$ combines attention modules, deep supervision, and ensemble learning across different input resolutions, convolutional block types, and loss functions. We also introduce $\mathit{ISLA\text{-}MT}$, a variant that leverages unlabeled target-domain images through semi-supervised learning. We trained both approaches on the SOOP and ISLES 2022 datasets and evaluated on a held-out internal validation set of 253 cases, as well as on 590 cases from an independent multicenter clinical cohort. We compared them with three state-of-the-art pre-trained models and with $\mathit{nnU\text{-}Net}$ trained on the same data. $\mathit{ISLA}$ outperformed the pre-trained models on the internal validation set, while both $\mathit{ISLA}$ and $\mathit{ISLA\text{-}MT}$ outperformed them on the external cohort. Although $\mathit{nnU\text{-}Net}$ performed slightly better than $\mathit{ISLA}$ internally, both $\mathit{ISLA}$ variants demonstrated better generalization to the external cohort. Furthermore, $\mathit{ISLA\text{-}MT}$ significantly improved performance over $\mathit{ISLA}$ on the external cohort. The proposed models performed particularly well on low-contrast lesions, and qualitative error analyses suggested greater robustness to acquisition artifacts. Both models are fast ($\sim$ 5 s per case), and their code and weights are publicly available at this https URL.
- [639] arXiv:2601.10455 (replaced) [pdf, html, other]
-
Title: Beyond Trajectory Similarity: Evaluating Surgical Planning by Protocol AdherenceSubjects: Computation and Language (cs.CL); Robotics (cs.RO)
Surgical planning is commonly evaluated by trajectory similarity to a single observed execution, a paradigm borrowed from activity anticipation benchmarks such as Ego4D. This conflates two distinct concepts: task completion and protocol adherence and assumes similarity to a completed surgical trajectory implies a valid plan. In surgery, this assumption is flawed: the clinically relevant target is protocol adherence rather than mere completion, and the reference trajectory is only one observed execution that may itself deviate from the protocol. We therefore introduce an evaluation framework grounded in expert-defined protocol rules that directly measures protocol adherence. Under this framework, similarity-based metrics systematically fail: they penalize protocol-adherent plans that deviate from the reference trajectory, while rewarding protocol-violating plans that remain superficially similar to it. Evaluating Video-LLMs further reveals two separable failure modes: unreliable video-to-state grounding and under-constrained planning. Neither is resolved by model scaling alone, whereas structural knowledge injection provides consistent gains. Our results motivate a shift from trajectory similarity toward protocol adherence as the evaluation standard for surgical planning.
- [640] arXiv:2601.15235 (replaced) [pdf, html, other]
-
Title: Tracing 3D Anatomy in 2D Strokes: Can Projection-Derived Masks Replace 3D Segmentation for Cervical Spine Fracture Recognition?Fabi Nahian Madhurja, Rusab Sarmun, Muhammad E. H. Chowdhury, Adam Mushtak, Israa Al-Hashimi, Sohaib Bassam ZoghoulComments: Includes supplementary material. Under revision at Medical Image AnalysisSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cervical spine fractures require rapid and accurate diagnosis, yet automatic CT interpretation remains challenging as subtle injuries must be assessed across large 3D volumes. We ask whether full 3D vertebra segmentation is necessary for fracture recognition, or whether vertebra masks approximated from 2D projections preserve sufficient diagnostic context. In the proposed multi-stage pipeline, a YOLOv8 detector localizes the cervical spine from multi-view variance projections (3D mean Intersection over Union 94.45%). Multi-label vertebra segmentation is performed with a DenseNet121-Unet on energy-based sagittal and coronal projections (mean Dice 87.86%). The predicted 2D masks are back-projected and fused into approximate 3D masks per vertebra to extract volumes of interest from the original CT, which are analyzed by an ensemble of 2.5D spatio-sequential CNN-Transformer models. On the full RSNA 2022 cohort at native prevalence (2,019 patients, five-fold cross-validation), the pipeline yields vertebra-level and patient-level F1-scores of 68.15% and 82.28%, area under the receiver operating characteristic curve of 91.63% and 90.95%, and area under the precision-recall curve (AUPRC) of 75.64% and 92.00%. With the classifier architecture fixed, replacing 3D-segmentation masks by projection-derived masks changed vertebra-level and patient-level AUPRC by +1.03 and +0.65 points, and the pipeline differed from the reproduced RSNA 2022 winning solution by -0.55 and +0.15 points. No AUPRC difference was statistically significant, and fold-wise 95% confidence intervals excluded vertebra-level losses larger than 1.46 and 2.81 points, respectively. Explainability and interobserver analyses examine reliability. Projection-based mask approximation is therefore a viable proxy for 3D vertebra segmentation in cervical fracture recognition.
- [641] arXiv:2601.15598 (replaced) [pdf, html, other]
-
Title: Ternary Spiking Neural Networks Enhanced by Complemented Neurons and Membrane Potential AggregationComments: 42 pages, 12 figures. Manuscript under reviewSubjects: Neural and Evolutionary Computing (cs.NE)
Spiking Neural Networks (SNNs) are promising energy-efficient models and powerful framworks of modeling neuron dynamics. However, existing binary spiking neurons exhibit limited biological plausibilities and low information capacity. Recently developed ternary spiking neuron possesses higher consistency with biological principles (i.e. excitation-inhibition balance mechanism). Despite of this, the ternary spiking neuron suffers from defects including iterative information loss, temporal gradient vanishing and irregular distributions of membrane potentials. To address these issues, we propose Complemented Ternary Spiking Neuron (CTSN), a novel ternary spiking neuron model that incorporates an learnable complemental term to store information from historical inputs. CTSN effectively improves the deficiencies of ternary spiking neuron, while the embedded learnable factors enable CTSN to adaptively adjust neuron dynamics, providing strong neural heterogeneity. Furthermore, based on the temporal evolution features of ternary spiking neurons' membrane potential distributions, we propose the Temporal Membrane Potential Regularization (TMPR) training method. TMPR introduces time-varying regularization strategy utilizing membrane potentials, furhter enhancing the training process by creating extra backpropagation paths. We validate our methods through extensive experiments on various datasets, demonstrating remarkable performance advances.
- [642] arXiv:2601.22585 (replaced) [pdf, html, other]
-
Title: HetCCL: Efficient LLM Training on Heterogeneous Vendor GPUsHeehoon Kim, Chris Jaehwan Lee, Taejeoung Kim, Jongwon Park, Jinpyo Kim, Pyongwon Suh, Ryan H. Choi, Sangwoo Lee, Jaejin LeeComments: Accepted at NeurIPS 2026Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
The rapid growth of large language models is driving organizations to expand their GPU clusters, often with GPUs from multiple vendors. However, current deep learning frameworks lack support for collective communication across heterogeneous GPUs, leading to inefficiency and higher costs. We present HetCCL, a collective communication library that unifies vendor-specific backends and enables RDMA-based communication across GPUs without requiring driver modifications. HetCCL introduces two novel mechanisms that enable cross-vendor communication while leveraging optimized vendor libraries, NVIDIA NCCL and AMD RCCL. Evaluations on a multi-vendor GPU cluster show that HetCCL matches NCCL and RCCL performance in homogeneous setups while uniquely scaling in heterogeneous environments, enabling practical, high-performance training with both NVIDIA and AMD GPUs without changes to existing deep learning applications.
- [643] arXiv:2601.23237 (replaced) [pdf, html, other]
-
Title: Applications of QR-based Vector-Valued Rational ApproximationSubjects: Numerical Analysis (math.NA); Mathematical Software (cs.MS)
Several applications of the QR-AAA algorithm, a greedy scheme for vector-valued rational approximation, are presented. The focus is on demonstrating the flexibility and practical effectiveness of QR-AAA in a variety of computational settings, including Stokes flow computation, multivariate rational approximation, function extension, the development of novel quadrature methods and near-field approximation in the boundary element method.
- [644] arXiv:2602.01953 (replaced) [pdf, html, other]
-
Title: Deep Multivariate Models with Parametric ConditionalsSubjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
We consider deep multivariate models for heterogeneous collections of random variables. In the context of computer vision, such collections may e.g. consist of images, segmentations, image attributes, and latent variables. When developing such models, most existing works start from an application task and design the model components and their dependencies to meet the needs of the chosen task. This has the disadvantage of limiting the applicability of the resulting model for other downstream tasks. Here, instead, we propose to represent the joint probability distribution by means of conditional probability distributions for each group of variables conditioned on the rest. Such models can then be used for practically any possible downstream task. Their learning can be approached as training a parametrised Markov chain kernel by maximising the data likelihood of its limiting distribution. This has the additional advantage of allowing a wide range of semi-supervised learning scenarios.
- [645] arXiv:2602.02557 (replaced) [pdf, html, other]
-
Title: The Alignment Curse: Modality Alignment Supercharges Audio Attacks via Text TransferComments: Accepted to NeurIPS 2026Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Sound (cs.SD)
Recent advances in end-to-end trained omni-models have substantially improved audio capabilities by strengthening text-audio modality alignment. However, whether such alignment inadvertently facilitates the transfer of safety vulnerabilities across modalities remains underexplored. This question is critical as text-based jailbreak attacks are considerably more mature than audio-based ones; if they transfer systematically, current audio safety evaluations may underestimate risks originating from the text modality. In this paper, we introduce the Alignment Curse, a formally characterized and empirically validated principle showing that stronger modality alignment enables more effective transfer of attacks from text to audio, revealing a fundamental tension between capability and safety. Motivated by this principle, we conduct a comprehensive black-box evaluation of three attack categories on recent omni-models (e.g., Qwen2.5-Omni, Qwen3-Omni): text attacks, text-transferred audio attacks, and audio attacks. We find that text-transferred audio attacks perform comparably to, and often better than, audio-based attacks, exhibiting a clear advantage under audio-only access. This suggests that text-based vulnerabilities play a pivotal role in shaping audio safety risks. Finally, we empirically analyze the relationship between modality alignment and transfer effectiveness across attack methods and models, observing consistent support for the Alignment Curse: tighter modality alignment leads to more effective cross-modality attack transfer.
- [646] arXiv:2602.05483 (replaced) [pdf, html, other]
-
Title: Toward Operationalizing Rasmussen in Evolving Software SystemsSubjects: Systems and Control (eess.SY); Computers and Society (cs.CY); Applications (stat.AP)
Rasmussen's qualitative drift model explains how successive, locally successful adaptations under efficiency pressures can move a system toward unacceptable operation. Brooker and Hochstein applied this perspective conceptually to software operations without specifying a quantitative procedure for observing drift. We take a step in this direction and propose a runtime observer that quantitatively operationalizes the model's observable part for evolving software. The proposed observer is intended to give engineers evidence for pausing or revising adaptation before declared operating limits are crossed. It tracks how activity or resource use is distributed across components and types of work. Measurement history quantifies accumulated change and recent movement toward or away from current limits. Assuming that direction and limits remain unchanged, the proposed observer calculates the first constraint ahead and the amount of further change required to reach it. Measurements remain comparable across supported component changes, while policy revisions are distinguished from changes in execution. Preliminary runtime studies demonstrated these distinctions under annotated software changes and revised retry limits. A separate multivariate exponentially weighted moving average (MEWMA) experiment detected CPU redistribution earlier with log-ratio coordinates than with raw shares, motivating investigation of representation-dependent observation. Further work will ground constraints in service outcomes, investigate runtime-model acquisition, and test whether decisions informed by observed drift reduce violations at acceptable useful-work loss and intervention cost.
- [647] arXiv:2602.06008 (replaced) [pdf, html, other]
-
Title: From Chats to Markets: AgenticPay for LLM-Powered Negotiation in Multi-Agent CommerceComments: Accepted at NeurIPS 2026 (Main Track)Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Large language model agents are increasingly expected to negotiate and transact autonomously, yet existing benchmarks rarely combine multimodal product grounding, multi-issue bargaining, and competitive multi-agent markets. We introduce AgenticPay, a unified benchmark for evaluating multimodal agents' ability to reach high-welfare, multidimensional agreements in realistic markets. AgenticPay includes 160 multimodal tasks across four business scenarios (E-commerce, Food Delivery, Ride-hailing, and Apartment Rental) and eight market topologies, from bilateral bargaining to many-to-many competition. Agents interpret product images, infer opponents' hidden preferences, and negotiate binding contract terms such as price, lease duration, return policy, and delivery speed through multi-round natural-language dialogue. A utility-based framework evaluates agreement quality and joint welfare. Experiments with proprietary (GPT-5.4, Claude Sonnet 4.6, Gemini 3.1 Pro Preview) and open-weight (Qwen3-VL-32B-Instruct, InternVL3-38B) multimodal models reveal substantial limitations in value creation and market reasoning. Even the strongest model, Gemini 3.1 Pro Preview, achieves a GlobalScore of only 42.3/100. As markets shift from bilateral to multi-sided, average GlobalScore falls by 5.7 points, with open-weight models declining by up to 11.5 points. AgenticPay provides a testbed for evaluating multimodal agents in realistic commerce. Code and data: this https URL.
- [648] arXiv:2602.09868 (replaced) [pdf, html, other]
-
Title: Free-GVC: Toward Retraining-Free Extreme Generative Video Compression with Temporal CoherenceComments: Accepted for publication in IEEE Transactions on BroadcastingSubjects: Computer Vision and Pattern Recognition (cs.CV)
Building on recent advances in video generation, generative video compression has emerged as a new paradigm for achieving visually pleasing reconstructions. However, existing methods exhibit limited exploitation of temporal correlations, causing noticeable flicker and degraded temporal coherence at ultra-low bitrates. In this paper, we propose Free-GVC, a retraining-free generative video compression framework that reformulates video coding as latent trajectory compression guided by a video diffusion prior. Our method operates at the group-of-pictures (GOP) level, encoding video segments into a compact latent space and progressively compressing them along the diffusion trajectory. To ensure perceptually consistent reconstruction across GOPs, we introduce an Adaptive Quality Control module that dynamically constructs an online rate-perception surrogate model to predict the optimal diffusion step for each GOP. In addition, an Inter-GOP Alignment module establishes frame overlap and performs latent fusion between adjacent groups, thereby mitigating flicker and enhancing temporal coherence. Experiments show that Free-GVC achieves an average of 93.29% BD-Rate reduction in DISTS over the recent neural codec DCVC-RT, and a user study further confirms its superior perceptual quality and temporal coherence at ultra-low bitrates.
- [649] arXiv:2602.11008 (replaced) [pdf, html, other]
-
Title: ROCKET: Rapid Optimization via Calibration-guided Knapsack Enhanced Truncation for Efficient Model CompressionAmmar Ali, Baher Mohammad, Denis Makhov, Dmitriy Shopkhoev, Magauiya Zhussip, Stamatios LefkimmiatisComments: Accepted at NeurIPS 2026Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
We present ROCKET, a training-free model compression method that achieves state-of-the-art performance in comparison with factorization, structured-sparsification and dynamic compression baselines. Operating under a global compression budget, ROCKET comprises two key innovations: First, it formulates layer-wise compression allocation as a multi-choice knapsack problem, selecting the optimal compression level for each layer to minimize total reconstruction error while adhering to a target model size. Second, it introduces a single-step sparse matrix factorization inspired by dictionary learning: using only a small calibration set, it sparsifies weight coefficients based on activation-weights sensitivity and then updates the dictionary in closed form via least squares bypassing iterative optimization, sparse coding, or backpropagation entirely. ROCKET consistently outperforms existing compression approaches across different model architectures at 20-50\% compression rates. Notably, it retains over 90\% of the original model's performance at 30\% compression without any fine-tuning. Moreover, when applying a light fine-tuning phase, recovery is substantially enhanced: for instance, compressing Qwen3-14B to an 8B-parameter model and healing it with just 30 million tokens yields performance nearly on par with the original Qwen3-8B. The code for ROCKET is at this http URL.
- [650] arXiv:2602.11454 (replaced) [pdf, html, other]
-
Title: Adaptive Power Iteration Method for Differentially Private PCASubjects: Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG)
We study $\left(\epsilon,\delta\right)$-differentially private algorithms for the problem of approximately computing the top singular vector of a matrix $A\in\mathbb{R}^{n\times d}$ where each row of $A$ is a data point in $\mathbb{R}^{d}$. Following Dwork-Talwar-Thakurta-Zhang (STOC 2014), we consider the privacy model where neighboring inputs differ by one single row. We give a novel algorithm that achieves beyond-worst-case guarantees for input matrices with low coherence, which is a structural property of matrices in many applications, including but not limited to i.i.d. data. Our algorithm contributes to the extensive literature on private power iteration methods, where we introduce a new filtering technique which adapts to this coherence parameter. Our work departs from and complements the work by Hardt-Roth (STOC 2013) which achieves beyond-worst-case guarantees for the more restrictive privacy model where neighboring inputs differ in one single entry by at most 1.
- [651] arXiv:2602.12683 (replaced) [pdf, html, other]
-
Title: Flow Matching from Viewpoint of Proximal OperatorsComments: 58 pages, 13 figuresSubjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
We discuss the dynamics of Optimal Transport Conditional Flow Matching (OT-CFM) under the manifold hypothesis, where the support of the target distribution is a low dimensional manifold. We first derive an exact proximal expression of the vector field, using an extended Brenier potential for the OT; the vector field can be expressed by a denoiser, which is presented as the proximal operator associated with the potential function. Based on this explicit proximal form of the vector field, we analyze the manifold stability for perturbations to initial points, using the Lyapunov exponents of the dynamics. By analyzing the local behavior with the tools in convex analysis, we show that the dynamics contract linearly in the perturbation of the normal directions to the manifold, while perturbations in the tangential directions are preserved; this implies the stability of the manifold structure to the perturbations. To justify the theoretical arguments on the population OT, we also establish the convergence of minibatch OT-CFM with neural network learning to population OT-CFM under regularity and stability assumptions, with rate exponents governed by the intrinsic dimension of the data manifold. Experiments on synthetic data, MNIST, and CIFAR-10 support the predicted geometric behavior.
- [652] arXiv:2602.15189 (replaced) [pdf, html, other]
-
Title: ScrapeGraphAI-100k: Dataset for Schema-Constrained LLM GenerationSubjects: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Producing output that conforms to a specified JSON schema underlies tool use, structured extraction, and knowledge base construction in modern large language models. Despite this centrality, public datasets for the task remain small, synthetic, or text-only, and rarely pair real page content with the prompts and schemas used in practice. We introduce ScrapeGraphAI-100k, 93,695 schema-constrained extraction events collected via opt-out ScrapeGraphAI telemetry in Q2--Q3 2025, deduplicated and balanced by schema from 9M raw events. The corpus spans 25,453 unique schemas, across 15 named languages, with English covering 89% of detected content. Each instance pairs Markdown-converted page content with a prompt, schema, LLM response, and per-example jsonschema-rs structural conformance labels. We characterize structural diversity across the corpus and identify sharp failure thresholds as schema complexity grows. As a case study, a 1.7B student fine-tuned on this data closely tracks the output distribution of its GPT-5-nano teacher, though it still trails a 30B-A3B reference (3.3B active parameters) on schema compliance. We offer this distillation result as preliminary evidence that grounding schema-constrained generation in real practitioner workloads at scale enables training and benchmarking that prior synthetic or text-only corpora could not support. ScrapeGraphAI-100k is publicly available on HuggingFace.
- [653] arXiv:2602.17544 (replaced) [pdf, html, other]
-
Title: Evaluating Chain-of-Thought Reasoning through Reusability and VerifiabilitySubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Information Retrieval (cs.IR)
AI applications increasingly compose multiple models, where one model acts on the output produced by another model. The transferred text can also carry the reasoning of the sender model. Chain-of-thought is our experimental instance of such text. We ask three questions about such inter-model reasoning transfer. Is the reasoning transmitted reliably? Is it useful to the receiver? Is it safe? We study these three questions with a Sender-Receiver setup. A Sender LLM produces a reasoning trace. A separate Receiver LLM must act on it. Neither model knows the identity of the other. The Sender therefore cannot tailor its reasoning to a particular Receiver. Reliability asks whether the Receiver recovers the Sender's intended conclusion. We measure it with Verifiability. Usefulness asks whether transferred reasoning helps a Receiver correct its mistakes. We measure it with Correctability. Safety asks whether transferred reasoning makes a Receiver abandon a correct answer and follow the wrong reasoning supplied by the Sender. We measure it with Deceptiveness. All three factors (Task, Sender, and Receiver) affect all three metrics that we measure. However, there are some interesting trends. Nature of task is a more important factor in the reliability of reasoning transfer. Reliability is higher on arithmetic tasks and lower on science and commonsense tasks. The nature of Receiver plays more important role while deciding the usefulness of reasoning transfer. The nature of the Sender matters most for whether corrupted reasoning harms the Receiver.
- [654] arXiv:2602.20659 (replaced) [pdf, html, other]
-
Title: Recursive Belief Vision Language Action ModelsComments: Accepted at the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)Subjects: Artificial Intelligence (cs.AI)
Vision-language-action models must enable agents to execute long-horizon tasks under partial observability. However, most existing approaches remain observation-driven, relying on short context windows or repeated queries to vision-language models (VLMs). This leads to loss of task progress, action repetition under perceptual aliasing, and high inference latency. While semantic grounding is important, long-horizon manipulation fundamentally requires persistent, action-conditioned state representations. Current VLAs lack such representations and exhibit limited temporal and physical reasoning, making them ill-suited for multi-stage control. This paper introduces RB-VLA, a belief-centric architecture trained with self-supervised world-model objectives that maintains a compact latent state encoding task-relevant history, dynamics, and object interactions. Queried once per task, the VLM provides high-level intent, while the belief tracks task progress and enables phase-aware, causally grounded control under partial observability without storing raw observations or scaling memory with time. The belief and intent jointly condition a diffusion policy for robust closed-loop execution. RB-VLA outperforms prior VLAs on long-horizon benchmarks, achieving 52.5 percent and 37.5 percent higher success rates on multi-stage pick-and-place and stacking tasks, respectively, compared to pi_0. It also reduces inference latency by up to five times relative to baselines and eliminates memory growth across timesteps observed in existing VLAs. Ablations show the belief module is the primary driver of performance, increasing success rates from 32.5 percent without belief to 77.5 percent with belief.
- [655] arXiv:2602.24195 (replaced) [pdf, html, other]
-
Title: Uncertainty Quantification for Multimodal Large Language Models with Incoherence-adjusted Semantic VolumeComments: Accepted at NeurIPS 2026. Earlier versions presented at ICLR 2025 QUESTION workshop and ICML 2025 R2-FM workshopSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Despite their capabilities, Multimodal Large Language Models (MLLMs) may produce plausible but erroneous outputs, hindering reliable deployment. Accurate uncertainty metrics could enable escalation of unreliable queries to human experts or larger models for improved performance. However, existing uncertainty metrics have practical constraints, such as being designed only for specific modalities, reliant on external tools, or computationally expensive. We introduce UMPIRE, a training-free uncertainty quantification framework for MLLMs that works efficiently across various input and output modalities without external tools, relying only on the models' own internal modality features. Inspired by Determinantal Point Processes (DPPs), UMPIRE computes a novel incoherence-adjusted semantic volume of sampled MLLM responses for a given task instance, which couples semantic embedding dispersion with the quadratic entropy of the responses. This approach effectively captures both the global semantic diversity of samples and the local incoherence of responses based on internal model confidence. We provide theoretical analysis motivating UMPIRE's design, and extensive experiments show that UMPIRE consistently outperforms baseline metrics in error detection and uncertainty calibration across image, audio, and video-text benchmarks, even with small sample budgets. We also demonstrate UMPIRE's generalization to non-text output tasks, including image and audio generation. Our code is available at this https URL.
- [656] arXiv:2603.00496 (replaced) [pdf, html, other]
-
Title: ESENSC: A Polynomial-Time Axiomatic Alternative to SHAPComments: Accepted to NeurIPS 2026 (camera-ready version). 28 pages. Title changed; the method previously called ESENSC_rev2 is renamed ESENSCSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
We propose ESENSC, a computationally efficient and theoretically grounded alternative to SHAP. ESENSC is a deterministic, closed-form attribution rule that requires only linearly many coalition-value evaluations, satisfies the null-player property, and admits an axiomatic characterization based on efficiency, a restricted form of differential marginality, and explicit computational this http URL, on tabular prediction tasks using neural network and XGBoost models, ESENSC shows close agreement with exact SHAP in both attribution magnitude and feature rankings while requiring substantially less computation. Under the evaluated settings, it compares favorably with widely used sampling-based SHAP approximations. While exact SHAP requires exponentially many coalition evaluations, ESENSC scales linearly with the number of features. These results show that ESENSC offers a favorable trade-off between agreement with exact SHAP and computational cost, providing a practical and theoretically grounded alternative for feature attribution in high-dimensional applications.
- [657] arXiv:2603.07932 (replaced) [pdf, html, other]
-
Title: Hard/Soft NLoS Detection via Combinatorial Data Augmentation for 6G PositioningSang-Hyeok Kim (1), Seung Min Yu (2), Jihong Park (3), Seung-Woo Ko (1) ((1) Inha University, South Korea, (2) Korea Railroad Research Institute, South Korea, (3) Singapore University of Technology and Design, Singapore)Comments: 24 pages, 15 figures, 13 tables. An earlier version of this work was presented in part at the IEEE Wireless Communications and Networking Conference (WCNC) 2026. This work has been submitted to the IEEE for possible publicationSubjects: Information Theory (cs.IT); Networking and Internet Architecture (cs.NI)
A key enabler for meeting the stringent requirements of 6G positioning is the ability to exploit site-dependent information governing line-of-sight (LoS) and non-line-of-sight (NLoS) propagation. However, acquiring such environmental information in real time is challenging in practice. To address this issue, we propose a novel NLoS detection algorithm termed combinatorial data augmentation-guided NLoS detection (CDA-ND), which builds upon our prior work. CDA-ND generates numerous preliminary estimated locations (PELs) by applying multilateration over many gNodeB (gNB) combinations using a single snapshot of range measurements. When a target gNB is in NLoS, the resulting PELs split into two clusters: one derived using the target gNB's range measurement and the other derived without it. Their displacement is summarized by a single vector, called the NLoS evidence vector (NEV), which is used to compute an NLoS likelihood score. Based on this score, two modes of NLoS detection are developed. First, each gNB is classified as LoS or NLoS, termed hard decision (HD), using a simple threshold test. Second, each gNB's NLoS confidence is probabilistically quantified, termed soft decision (SD), which extends HD with weak site-survey priors, namely empirical NLoS-score samples and the average NLoS probability. We then design positioning algorithms tailored to these two modes by excluding gNBs deemed NLoS and re-weighting the remaining gNBs for SD. The proposed CDA-ND achieves high reliability in indoor factory environments under frequency range 1, attaining NLoS detection accuracies of 96.8% and 91.5% when the proportion of NLoS gNBs is approximately 18% and 56%, respectively. As a result, integrating CDA-ND into positioning significantly reduces mean absolute error by about 22% and 67% in LoS- and NLoS-dominant environments, respectively.
- [658] arXiv:2603.08036 (replaced) [pdf, html, other]
-
Title: Samyama: A Unified Graph-Vector Database with In-Database Optimization, Agentic Enrichment, and Hardware AccelerationComments: 20 pages, 4 figures. v4 adds Section 7.6 on synthetic workload compilation and privacy-preserving graph twins, and corrects the v1-v2 hardware history noteSubjects: Databases (cs.DB)
Modern data architectures fragment across graph databases, vector stores, analytics engines and optimization solvers, forcing ETL between them. We present Samyama, a graph-vector database in Rust that unifies these workloads in one engine: a RocksDB-backed store with MVCC, a vectorized executor, a cost-based planner, a CSR analytics engine, RDF and SPARQL, 22 metaheuristic solvers callable from the query language, HNSW vector indexing, and agentic enrichment that expands a graph via LLMs. It has been run to billion-edge scale: 74.3M nodes and 1.07B edges from four biomedical sources on one machine for $2.50 of spot compute. This version re-measures the system at release v1.8.0 on rented Linux hosts a reader can boot, replacing earlier figures taken on a Mac Mini that nobody could re-run. On a 16-vCPU cloud instance, ingestion reaches 123K-152K nodes/s and PageRank costs 6.7 ms per iteration at ten thousand nodes and 236 ms at a million. The previously reported 8.2x GPU speedup is withdrawn; in its place we measure the CUDA path, which no earlier version measured at all, on a rented NVIDIA A16: parity with the CPU at ten thousand nodes, 1.62x at a hundred thousand and 2.70x at a million, with 1.1-1.35 s of context initialisation on the first call. The earlier Cypher-throughput figure, its claim of near-constant index-driven scaling, and a vector-search figure are withdrawn as unreproducible. openCypher conformance, once estimated at ~90%, is now measured at 99.9% of evaluated TCK scenarios. Against Neo4j 5 and FalkorDB on one host over an identical SF10 extract, 14 of 14 SNB complex reads land within 5x of the best competitor and 7 of 14 are faster.
- [659] arXiv:2603.09424 (replaced) [pdf, html, other]
-
Title: System-wide Dynamic Performance Metric for Converter-Dominated Power SystemsSubjects: Systems and Control (eess.SY)
Fast controller dynamics in Inverter-Based Resources (IBRs)-dominated power networks couple frequency and voltage, making traditional decoupled analysis inadequate. We introduce a unified system-wide metric---the complex frequency of system losses-to evaluate dynamic performance and synchronization. Composed of device- and network-driven terms, it expresses voltage and injected-current phasor variations weighted by bus complex powers. Validated on an IBR-modified IEEE 39-bus system and the Irish transmission grid, the framework effectively assesses R/X sensitivities, dynamic support limits, and data-center load impacts.
- [660] arXiv:2603.09527 (replaced) [pdf, html, other]
-
Title: Efficiently Aligning Draft Models via Parameter- and Data-Efficient AdaptationLuxi Lin, Zhihang Lin, Zhanpeng Zeng, Yuhao Chen, Qingyu Zhang, Jixiang Luo, Xuelong Li, Rongrong JiComments: 10 pagesSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Speculative decoding accelerates LLM inference but suffers from performance degradation when target models are fine-tuned for specific domains. A naive solution is to retrain draft models for every target model, which is costly and inefficient. To address this, we introduce a practical parameter- and data-efficient framework named Efficient Draft Adaptation, abbreviated as EDA, for efficiently adapting draft models. EDA introduces three innovations: (1) a decoupled architecture that utilizes shared and private components to model the shared and target-specific output distributions separately, enabling parameter-efficient adaptation by updating only the lightweight private component; (2) a data regeneration strategy that utilizes the fine-tuned target model to regenerate training data, thereby improving the alignment between training and speculative decoding, leading to higher average acceptance length; (3) a sample selection mechanism that prioritizes high-value data for efficient adaptation. Our experiments show that EDA effectively restores speculative performance on fine-tuned models, achieving superior average acceptance lengths with significantly reduced training costs compared to full retraining. Code is available at this https URL.
- [661] arXiv:2603.10099 (replaced) [pdf, html, other]
-
Title: Denoising the US Census: Succinct Block Hierarchical RegressionComments: This is the author's version of the work. It is posted here by permission of the AAAS for personal use, not for redistribution. The definitive version was published in Science on November 5, 2026, DOI: https://doi.org/10.1126/science.aej1765. The final published version is available at this https URLJournal-ref: Science 394, 686-692 (2026)Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
The 2020 United States Census adopted differential privacy to protect individual confidentiality, adding calibrated noise to summary demographic measurements that are then reconciled into population tables. We developed BlueDown, a postprocessing algorithm that combines techniques from linear regression and constrained optimization while exploiting measurement structure for efficiency. On 2020 Census data, BlueDown reduced the error of groupings by race at the county level by 21-36% and of all county-level groupings by 9-58%; at the tract level, it reduced error of groupings by housing type by 38-61%. These gains came at no cost to confidentiality or computational efficiency, and would improve the validity of research relying on this definitive dataset and analyses used to guide the distribution of over $1.5 trillion in annual federal spending.
- [662] arXiv:2603.14667 (replaced) [pdf, html, other]
-
Title: Comparative Analysis of 3D Convolutional and 2.5D Slice-Conditioned U-Net Architectures for MRI Super-Resolution via Elucidated Diffusion ModelsComments: v2: revised version resubmitted to IEEE Access; supersedes the results of v1. 10 pages, 8 figures, 6 tablesSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Magnetic resonance imaging (MRI) super-resolution (SR) methods that computationally enhance low-resolution acquisitions offer a software alternative to higher-field scanners. We study an elucidated diffusion model (EDM) framework for 2x in-plane brain MRI SR and compare two U-Net backbones: (i) a 3D convolutional U-Net that processes volumetric patches and (ii) a 2.5D slice-conditioned U-Net that super-resolves each slice while conditioning on its two neighbouring slices. All models are trained from scratch to convergence on 500 subjects of the NKI cohort of FOMO60K, with a separate 6-subject validation set for checkpoint selection, and evaluated on 70 held-out subjects with subject-level statistics against baselines (EDSR, Swin2SR, Res-SRDiff) trained on the same data. Neither architecture is better overall. On the degradation used in training the two are tied (37.72 dB PSNR for the 3D model and 37.70 dB for the 2.5D model), 0.2 dB ahead of EDSR. When the test degradation changes to k-space truncation, EDSR matches the diffusion models, and with added noise the 2.5D model is the most accurate method. When the test cohort changes to 70 subjects of an external dataset (IXI), the 3D model is the most accurate, 0.4 dB ahead of the best baseline and 0.9 dB ahead of the 2.5D model in foreground PSNR.
- [663] arXiv:2603.15854 (replaced) [pdf, html, other]
-
Title: FlashSampling: Fast and Memory-Efficient Exact SamplingComments: Project Page: this https URLSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Sampling from a categorical distribution is mathematically simple, but in large-vocabulary decoding, it often triggers extra memory traffic and extra kernels after the LM head. We present FlashSampling, an exact sampling primitive that fuses sampling into the LM-head matmul and never materializes the logits tensor in HBM. The method is simple: compute logits tile-by-tile on chip, add Gumbel noise, keep only one maximizer per row and per vocabulary tile, and finish with a small reduction over tiles. In tensor-parallel decoding, FlashSampling replaces the all-gather of logits with streaming peer-to-peer writes: This overlaps GPU-to-GPU communication with computation and HBM loads across up to 8 GPUs, with near-ideal scaling at large batch sizes. Our kernel is exact because argmax decomposes over partitions; grouped variants for online and tensor-parallel settings are exact by hierarchical factorization of the categorical distribution. FlashSampling demonstrates kernel-level speedups on decode workloads across 4 different datacenter GPUs (H100, H200, B200, B300), and in end-to-end vLLM experiments, it reduces time per output token by up to $10%$ on the models we test. These results show that exact sampling, with no approximation, can be integrated into the matmul itself, consolidating the bandwidth-bound sampling step in an efficient epilogue.
- [664] arXiv:2603.16411 (replaced) [pdf, html, other]
-
Title: RECOVER: Robust Entity Correction via agentic Orchestration of hypothesis Variants for Evidence-based RecoveryComments: Accepted to EMNLP 2026 Industry TrackJournal-ref: Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages 2939-2951, Budapest, HungarySubjects: Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
Entity recognition in Automatic Speech Recognition (ASR) is challenging for rare and domain-specific terms. In domains such as finance, medicine, and air-traffic control, these errors are costly. If an entity is entirely absent from the ASR output, text-only post-correction has little signal to recover it. To address this, we introduce RECOVER, an agentic post-ASR entity-correction framework. It leverages multiple hypotheses as evidence from ASR, retrieves relevant entities, proposes Large Language Model (LLM) correction under constraints, filters them through deterministic guardrails and uses an LLM critic that judges each surviving edit and can trigger a targeted feedback round to recover an entity the first pass missed. The hypotheses are used by one of several strategies, namely 1-Best, Entity-Aware Select, Recognizer Output Voting Error Reduction (ROVER) Ensemble, LLM-Select and Adaptive Routing. Evaluated across five diverse datasets over a Whisper-large-v3 baseline, it achieves up to 49% relative reduction in entity-phrase word error rate (E-WER) and increases entity recall by up to 14 percentage points while preserving the overall word error rate (WER), with LLM-Select the strongest strategy.
- [665] arXiv:2603.16840 (replaced) [pdf, other]
-
Title: What DINO saw: ALiBi positional encoding reduces positional bias in Vision TransformersMoritz Pawlowsky, Antonis Vamvakeros, Alexander Weiss, Anja Bielefeld, Samuel J. Cooper, Ronan DochertySubjects: Computer Vision and Pattern Recognition (cs.CV); Materials Science (cond-mat.mtrl-sci)
Vision transformers (ViTs) - especially feature foundation models like DINOv2 - learn rich representations useful for many downstream tasks. However, architectural choices (such as positional encoding) can lead to these models displaying positional biases and artefacts independent of semantic content. This makes zero-shot adaption difficult in fields like material science, where images are often cross-sections of homogeneous microstructure (i.e. having no preferred direction). In this work, we investigate the positional bias in ViTs via linear probing, finding it present across a range of objectives and positional encodings, and subsequently reduce it by finetuning models to use ALiBi relative positional encoding. We demonstrate that these models retain desirable general semantics and their unbiased features can be used successfully in trainable segmentation of complex microscopy images.
- [666] arXiv:2603.21465 (replaced) [pdf, html, other]
-
Title: DRTriton: Large-Scale Synthetic Data Driven Reinforcement Learning for Triton Kernel GenerationComments: Accepted at NeurIPS 2026. Code available at this https URLSubjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Developing efficient CUDA kernels is a fundamental yet challenging task in the generative AI industry. Recent research leverages Large Language Models (LLMs) to automatically convert PyTorch reference implementations to CUDA kernels, significantly reducing engineering effort. State-of-the-art LLMs, such as GPT-5.2 and Claude-Sonnet-4.5, still struggle with this task. To address this challenge, we propose DRTriton, a scalable learning framework for training LLMs to convert PyTorch programs into highly optimized Triton kernels, which are then compiled to CUDA kernels at runtime. DRTriton consists of three key components: (i) a data synthetic algorithm CSP-DAG that guarantees full coverage and unbiased uniform sampling over the operator space with controlled difficulty; (ii) a curriculum RL framework with decoupled rewards that jointly optimizes conversion success rate and execution speed; and (iii) a test-time search algorithm that further improves the execution speed of the generated Triton kernels. With a warmup stage of SFT on limited PyTorch-Triton pairs curated using existing LLMs, DRTriton trained by RL on synthesized PyTorch programs generalizes effectively to real-world CUDA kernels that are challenging even for human experts. Experimental results show that DRTriton-7B achieves speedup over PyTorch on 92% of KernelBench Level 2 tasks, compared to 23% for GPT-5.2 and 19% for Claude-Sonnet-4.5.
- [667] arXiv:2603.26378 (replaced) [pdf, html, other]
-
Title: Generative machine learning for protein designSenura Hansaja Wanasekara, Manoj M Wagle, Minh-Duong Nguyen, Xiaochen Liu, Nguyen H. Tran, Manolis Kellis, Ken-Tye YongComments: v2: Substantial revision. The paper has been restructured and now proceeds as introduction, approaches to generative modeling, representations for protein design, designing proteins with generative models, evaluating generative protein design, and challenges and future directionsSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Designing proteins with specific biological functions is of considerable interest for both the development of new therapeutics and the engineering of enzymes, yet achieving the intended activity remains challenging. Generative artificial intelligence (AI) has emerged as a powerful tool to explore the vast space of possible designs by learning from protein sequence and structural data. Here, we review 59 generative approaches across six protein design tasks, including sequence generation, backbone generation, structure-conditioned sequence design, joint sequence-structure design, motif scaffolding, and binder and antibody design. We outline common generative approaches and protein representations and discuss how structural constraints and functional information are utilized during generation. This is followed by a discussion of the computational and experimental approaches used to evaluate designed proteins and what such assessments reveal about their properties. Finally, we discuss current limitations in designing proteins beyond well-represented families and incorporating experimental feedback into the design process. We expect that rapid advances in generative modeling will enable more reliable design of a wider range of protein structures and functions over the coming years.
- [668] arXiv:2604.01171 (replaced) [pdf, html, other]
-
Title: Open-Set Supervised 3D Anomaly Detection: An Industrial Dataset and a Generalisable Framework for Unknown DefectsHanzhe Liang, Luocheng Zhang, Junyang Xia, HanLiang Zhou, Bingyang Guo, Yingxi Xie, Can Gao, Ruiyun Yu, Jinbao Wang, Pan LiSubjects: Computer Vision and Pattern Recognition (cs.CV)
Although self-supervised 3D anomaly detection assumes that acquiring high-precision point clouds is computationally expensive, in real manufacturing scenarios it is often feasible to collect a limited number of anomalous samples. Therefore, we study open-set supervised 3D anomaly detection, where the model is trained with only normal samples and a small number of known anomalous samples, aiming to identify unknown anomalies at test time. We present Open-Industry, a high-quality industrial dataset containing 20 categories, each with five real anomaly types collected from production lines. We first adapt general open-set anomaly detection methods to accommodate 3D point cloud inputs better. Building upon this, we propose Open3D-AD, a point-cloud-oriented approach that leverages normal samples, simulated anomalies, and partially observed real anomalies to model the probability density distributions of normal and anomalous data. Then, we introduce a simple Correspondence Distributions Subsampling to reduce the overlap between normal and non-normal distributions, enabling stronger dual distributions modeling. Based on these contributions, we establish a comprehensive benchmark and evaluate the proposed method extensively on Open-Industry as well as established datasets including Real3D-AD and Anomaly-ShapeNet. Benchmark results and ablation studies demonstrate the effectiveness of Open3D-AD and further reveal the potential of open-set supervised 3D anomaly detection.
- [669] arXiv:2604.06463 (replaced) [pdf, html, other]
-
Title: A Control Barrier Function-Constrained Model Predictive Control Framework for Safe Reinforcement LearningComments: 10 pages; additional theoretical results and expanded discussionSubjects: Systems and Control (eess.SY)
Ensuring safety under unknown and stochastic dynamics remains a significant challenge in reinforcement learning (RL). In this paper, we propose a model predictive control (MPC)-based safe RL framework, called Probabilistic Ensembles with CBF-constrained Trajectory Sampling (PECTS), to address this challenge. PECTS jointly learns stochastic system dynamics with probabilistic neural networks (PNNs) and control barrier functions (CBFs) with Lipschitz-bounded neural networks. Safety is enforced by incorporating learned CBF constraints into the MPC formulation while accounting for the model stochasticity. This enables probabilistic safety under model uncertainty. To solve the resulting MPC problem, we utilize a sampling-based optimizer together with a safe trajectory sampling method that discards unsafe trajectories based on the learned system model and CBF. We also derive a finite-horizon exit-probability bound using tail information for centered barrier increments. We validate PECTS in various simulation studies, where it outperforms baseline methods.
- [670] arXiv:2604.06822 (replaced) [pdf, html, other]
-
Title: Non-RS type cyclic MDS codes over finite fields via cyclotomic field reductionComments: 17pagesSubjects: Information Theory (cs.IT)
Cyclic maximum distance separable (MDS for short) codes are a special subclass of linear codes and have received a lot of attention, as these codes have very important applications in many areas including quantum codes, designs and finite geometry. However, the existing construction methods for cyclic MDS codes mainly rely on strict restrictions on certain parameters or involve relatively complex procedures. In this paper, we investigate this problem further via norm reduction in cyclotomic fields. By converting the verification of the MDS property over a finite field into checking non-zero minors in characteristic zero, we propose a construction method of cyclic MDS codes over finite fields via cyclotomic field reduction. Based on this method, we obtain several cyclic MDS codes over finite fields and many non-RS type cyclic MDS codes are produced. Compared with the existing construction method in [Non-Reed-Solomon Type Cyclic MDS Codes, IEEE Trans. Inf. Theory, 71(5): 3489--3496, 2025], we give a new and relatively direct characterization method for constructing cyclic MDS codes over finite fields and the obtained non-RS cyclic MDS codes have flexible parameters in terms of the code length $n$, the defining set $J$, and the underlying finite field $\mathbb{F}_q$.
- [671] arXiv:2604.07055 (replaced) [pdf, html, other]
-
Title: Resolving AdaBoost Cycling with LLMs: A Computer-Assisted CounterexampleSubjects: Machine Learning (cs.LG)
We develop a computer-assisted counterexample to the open question of whether exhaustive AdaBoost always converges to a finite cycle, posed by Rudin, Schapire, and Daubechies in COLT 2012. The construction is based on a block-product of two small AdaBoost instances whose two factors share an exact period-2 orbit for their five-step branch maps, but whose linearized return maps have dominant eigenvalues with an irrational logarithmic ratio. This irrationality forces the sequence of instances that win each round of updates to have an irrational asymptotic frequency, which precludes convergence to a finite cycle. The resulting counterexample has 2,500 examples, weak-learning margin $1/5$, and a uniform initial distribution, and it holds under every maximizing tie rule. All finite computations in the proof are certified by rational arithmetic in SymPy, with SageMath verifying the main inequalities. We also document the collaborative workflow that led to the construction, in which GPT-5.4 Pro produced the key technical arguments; Claude Opus 4.6 summarized and critiqued different versions of that work, proposing new directions to pursue; and the authors orchestrated the process by selecting directions to prioritize, resolving ambiguities, refining each argument, and verifying the final solution. Later conversations with GPT-5.6 Sol and GPT-6 Astra substantially simplified the solution. We present this as one of the first in-depth case studies of LLM-assisted mathematical research, resolving a long-standing open problem from theoretical machine learning, and detail the advantages and challenges of such research.
- [672] arXiv:2604.07953 (replaced) [pdf, html, other]
-
Title: Pruning Extensions and Efficiency Trade-Offs for Sustainable Time Series ClassificationRaphael Fischer, Angus Dempster, Sebastian Buschjäger, Matthias Jakobs, Urav Maniar, Geoffrey I. WebbSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Time series classification (TSC) at scale requires balancing predictive performance against resource consumption, which is subject to the given data, model, and hardware. This paper introduces a multi-dimensional evaluation framework that allows for explicitly benchmarking and exploring the resource-versus-quality trade-offs in TSC. To improve efficiency, we present a pruning strategy with bounded error for leading hybrid classifiers - Hydra and Quant - and moreover propose Hydrant, a novel extension that integrates interval representations with random convolutions and dictionary learning. With over 4000 evaluated configurations across 20 MONSTER datasets, 10 methods, and three compute setups, we systematically analyze how model design, hyperparameters, and hardware choices impact practical TSC performance. Our results show that pruning can significantly reduce energy consumption by up to 70% while only marginally lowering predictive quality. As such, our contributed evaluation framework, methodological extensions, experimental results, and software repository advance TSC toward sustainable and reproducible practice.
- [673] arXiv:2604.14397 (replaced) [pdf, html, other]
-
Title: Generating Concept Lexicalizations via Dictionary-Based Cross-Lingual Sense ProjectionComments: Paper presented at Canadian AI 2026Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
We study the task of automatically expanding WordNet-style lexical resources to new languages through sense generation. We generate senses by associating target-language lemmas with existing lexical concepts via semantic projection. Given a sense-tagged English corpus and its translation, our method projects the annotated synsets onto aligned target-language tokens and assigns the corresponding lemmas to those synsets. To generate alignments and ensure their quality, we augment a pretrained base aligner with a bilingual dictionary, which is also used to filter incorrect sense projections. We evaluate the method on multiple languages, comparing it to prior methods, as well as dictionary-based and large language model baselines. Results show that the proposed project-and-filter strategy improves precision while remaining interpretable and resource- efficient. We release our code, documentation, and generated sense inventories at this https URL.
- [674] arXiv:2604.14586 (replaced) [pdf, html, other]
-
Title: CPGRec+: A Balance-oriented Framework for Personalized Video Game RecommendationsComments: Published in ACM Transactions on Information Systems (TOIS). 43 pages, 9 figures. Added references on multi-view clusteringJournal-ref: ACM Trans. Inf. Syst. 44, 3, Article 66 (March 2026), 44 pagesSubjects: Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
The rapid expansion of gaming industry requires advanced recommender systems tailored to its dynamic landscape. Existing Graph Neural Network (GNN)-based methods primarily prioritize accuracy over diversity, overlooking their inherent trade-off. To address this, we previously proposed CPGRec, a balance-oriented gaming recommender system. However, CPGRec fails to account for critical disparities in player-game interactions, which carry varying significance in reflecting players' personal preferences and may exacerbate over-smoothness issues inherent in GNN-based models. Moreover, existing approaches underutilize the reasoning capabilities and extensive knowledge of large language models (LLMs) in addressing these limitations. To bridge this gap, we propose two new modules. First, Preference-informed Edge Reweighting (PER) module assigns signed edge weights to qualitatively distinguish significant player interests and disinterests while then quantitatively measuring preference strength to mitigate over-smoothing in graph convolutions. Second, Preference-informed Representation Generation (PRG) module leverages LLMs to generate contextualized descriptions of games and players by reasoning personal preferences from comparing global and personal interests, thereby refining representations of players and games. Experiments on \textcolor{black}{two Steam datasets} demonstrate CPGRec+'s superior accuracy and diversity over state-of-the-art models. The code is accessible at this https URL.
- [675] arXiv:2604.14833 (replaced) [pdf, html, other]
-
Title: Federated User Behavior Modeling for Privacy-Preserving LLM RecommendationSubjects: Information Retrieval (cs.IR)
Large Language Models have shown great success in recommender systems. However, the limited and sparse nature of user data often restricts the LLM's ability to effectively model behavior patterns. To address this, existing studies have explored cross-domain solutions by conducting Cross-Domain Recommendation tasks. But previous methods typically assume domains are overlapped and can be accessed readily. None of the LLM methods address the privacy-preserving issues in the CDR settings, that is, Privacy-Preserving Cross-Domain Recommendation. Conducting non-overlapping PPCDR with LLM is challenging since: 1)The inability to share user identity or behavioral data across domains impedes effective cross-domain alignment. 2)The heterogeneity of data modalities across domains complicates knowledge integration. 3)Fusing collaborative filtering signals from traditional recommendation models with LLMs is difficult, as they operate within distinct feature spaces. To address the above issues, we propose SF-UBM, a Semantic-enhanced Federated User Behavior Modeling method. Specifically, to deal with Challenge 1, we leverage natural language as a universal bridge to connect disjoint domains via a semantic-enhanced federated architecture. Here, text-based item representations are encrypted and shared, while user-specific data remains local. To handle Challenge 2, we design a Fact-counter Knowledge Distillation module to integrate domain-agnostic knowledge with domain-specific knowledge, across different data modalities. To tackle Challenge 3, we project pre-learned user preferences and cross-domain item representations into the soft prompt space, aligning behavioral and semantic spaces for effective LLM learning. We conduct extensive experiments on three pairs of real-world domains, and the experimental results demonstrate the effectiveness of SF-UBM compared to the recent SOTA methods.
- [676] arXiv:2604.17422 (replaced) [pdf, html, other]
-
Title: Where to Focus: Query-Modulated Multimodal Keyframe Selection for Long Video UnderstandingComments: 10 pages, 5 figures. The 34th ACM International Conference on Multimedia (MM '26)Journal-ref: Proceedings of the 34th ACM International Conference on Multimedia (MM '26), pp. 10508-10517, 2026Subjects: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
Long video understanding remains a formidable challenge for Multimodal Large Language Models (MLLMs) due to the prohibitive cost of processing dense frame sequences. Prevailing keyframe-selection methods rely on either a single visual-centric metric (e.g., CLIP similarity) or a static fusion of heuristic scores. This "one-size-fits-all" paradigm frequently fails: visual-only metrics are ineffective for plot-driven narrative queries, while indiscriminately adding textual scores introduces severe "modal noise" for purely visual tasks. To break this bottleneck, we propose Q-Gate, a plug-and-play, training-free framework that treats keyframe selection as a dynamic modality routing problem. We decouple retrieval into three lightweight expert streams: Visual Grounding for local details, Global Matching for scene semantics, and Contextual Alignment for subtitle-driven narratives. Crucially, Q-Gate introduces a Query-Modulated Gating Mechanism that uses the in-context reasoning of an LLM to assess query intent and dynamically allocate weights across the experts, activating necessary modalities while "muting" irrelevant ones to maximize the signal-to-noise ratio. Extensive experiments on LongVideoBench and Video-MME across multiple MLLM backbones show that Q-Gate outperforms representative keyframe-selection baselines in most settings, with particularly strong gains on long and medium videos, providing a robust and interpretable solution for scalable video reasoning. Code is available at this https URL
- [677] arXiv:2604.17714 (replaced) [pdf, html, other]
-
Title: Screen Before You Interpret: A Portable Validity Screen for Benchmark-Based LLM Confidence SignalsComments: v2: Corrected and consolidated; supersedes arXiv:2604.17707 and 2604.17716. Fixes errors in v1: declines were scored as confident errors, a free-text track was mis-graded, tiers did not come from the released tool, and the Yang et al. reanalysis was not computed from their data. 16 pages, 4 figures. Code and data: this https URLSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
LLM confidence signals drive abstention, routing and escalation, yet evaluations rarely check whether a signal varies with correctness before computing metrics on it. Clinical personality assessment handles the analogous problem by screening response validity before interpreting substantive scales. We specify a portable screen in that tradition. From the correctness-by-confidence table it computes the rate of high confidence on errors (L), the rate of low confidence on correct answers (Fp), and their contrast (RBS, equal to minus Youden's J). It adds no new statistics: it separates blanket confidence, blanket withdrawal and inversion, attaches interval guards, and calls a signal Valid only when it measurably tracks correctness. On 20 frontier models answering 452 two-alternative items with binary KEEP/WITHDRAW probes, one model is Invalid (DeepSeek-R1 withdraws 95% of its correct answers), four are Indeterminate and 15 are Valid. Two Indeterminate models keep confidence on 96% of their errors but make too few errors to certify it; certification needs 35 to 139 errors. Only 6 of the 20 verdicts reproduce reliably on half the items. On verbalized 0-100 confidence (33 models on 1,500 MMLU items, 20 on a second 500-item sample, and Yang et al.'s data for 11 models on 10 benchmarks), the verdict depends on the cut point: at 50 most models are Invalid because answers pile at the ceiling, while at 90-95 most pass. Battery item sensitivity does not predict MMLU discrimination (rho = -.07). Protocol amendments were settled on these data, so the verdicts are descriptive pending a held-out test. This version corrects scoring, grading, classification and reporting errors in version 1 and supersedes arXiv:2604.17707 and 2604.17716.
- [678] arXiv:2604.21909 (replaced) [pdf, html, other]
-
Title: Directional Confusions Reveal Divergent Inductive Biases Through Rate-Distortion Geometry in Human and Machine VisionSubjects: Computer Vision and Pattern Recognition (cs.CV); Information Theory (cs.IT); Neurons and Cognition (q-bio.NC)
Humans and modern vision models can reach similar classification accuracy while making systematically different kinds of mistakes -- differing not in how often they err, but in who gets mistaken for whom and in which direction. We show that these directional confusions reveal distinct inductive biases that are invisible to accuracy alone. Using matched human and deep vision model responses on a natural-image categorization task under 12 perturbation types, we quantify asymmetry in confusion matrices and link it to the shape of the information-error trade-off. We characterize this trade-off geometry using three signatures extracted from the rate-distortion (RD) frontier: slope beta, curvature kappa and efficiency (AUC). We find that humans exhibit broad but weak asymmetries, whereas deep vision models show sparser, stronger directional collapses. Robustness training reduces global asymmetry but fails to recover the human-like breadth-strength profile of graded similarity. Mechanistic simulations further show that different asymmetry organizations shift the RD frontier in opposite directions, even when matched for performance. Together, these results position directional confusions and RD geometry as compact, interpretable signatures of inductive bias under distribution shift.
- [679] arXiv:2604.26701 (replaced) [pdf, html, other]
-
Title: Explicit Planar Finite Element Elasticity Complexes and $C^1$ Elements on Barycentric RefinementsComments: 22 pages, 4 figuresSubjects: Numerical Analysis (math.NA)
We give a closed-form realization of the finite element elasticity complex associated with the Arnold--Douglas--Gupta family on barycentric refinements. Their construction implies the existence of minimal $C^1$ potential spaces, but does not provide explicit general-degree scalar enrichments. On each macrotriangle, we derive three divergence-free stress enrichments and their Airy potentials. For every degree $q\geq4$, the resulting potential space enriches $\mathbb P_q$ by three explicit functions. A geometric decomposition yields unisolvent degrees of freedom and $C^1$ assembly without vertex supersmoothness. The space attains the dimension bound for macroelements containing $\mathbb P_q$ with full polynomial boundary trace freedom. Its Airy image is the divergence-free subspace of the corresponding stress space, and global assembly gives an exact discrete elasticity complex on simply connected domains. The low-order spaces $U_2$ and $U_3$ recover the reduced and full Hsieh--Clough--Tocher spaces, respectively.
- [680] arXiv:2604.26741 (replaced) [pdf, html, other]
-
Title: Analytically Characterized Optimal Power Control for Signal-Level-Integrated Sensing, Computing and Communication in Federated LearningJournal-ref: IEEE Transactions on Wireless Communications, vol. 25, pp. 25941-25958, 2026Subjects: Information Theory (cs.IT); Signal Processing (eess.SP); Optimization and Control (math.OC)
In next-generation wireless edge systems, efficient functionality integration is essential to address the growing demands of communication, computation, and sensing. Signal-level integrated sensing, computing, and communication (Sig-ISCC) is envisioned, where a single waveform simultaneously supports sensing, computing and communication via over-the-air computation (AirComp). Meanwhile, federated learning (FL) is widely regarded as a promising distributed machine learning framework that enables network intelligence in a privacy-preserving and secure manner, and exhibits strong synergy with AirComp, which alleviates the communication bottleneck of FL. In this paper, we study uplink Sig-ISCC design for AirComp-FL with joint target detection. We formulate the joint power and receive-scaling control problem, where edge devices' transmitted signals should serve both sensing and AirComp purposes. The goal is to minimize the AirComp aggregation distortion subject to a joint target-detection requirement. Although the resulting problem is non-convex in the original variables, we show that it admits an equivalent convex reformulation after a suitable variable transformation. By exploiting analytical optimality properties, we develop a robust, optimal, and polynomial-time-complexity algorithm that efficiently achieves the optimal transmit powers and receive scaling factor. Simulation results validate the optimality and numerical robustness of the proposed algorithm and show its superior FL performance compared to baseline methods.
- [681] arXiv:2604.27308 (replaced) [pdf, html, other]
-
Title: BoostLoRA: Growing Effective Rank by Boosting AdaptersComments: PreprintSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Parameter-efficient fine-tuning (PEFT) methods face a tradeoff between adapter size and expressivity: ultra-low-parameter adapters are confined to fixed low-rank subspaces, capping performance even with extended training. We propose BoostLoRA, a gradient-boosting framework that overcomes this limit by iteratively training and merging minimal adapters on the examples the current model gets wrong. A ROTATE SVD basis strategy assigns each round to an orthogonal subspace, so cumulative effective rank grows linearly with the number of rounds while each adapter remains ultra-low-rank. After merging, adapters are discarded, leaving zero inference overhead. On Qwen2.5-3B, BoostLoRA reaches 89.1% on GSM8K and 68.8% on MATH-500, surpassing both the best single-shot ultra-low parameter adapter (TinyLoRA) and full fine-tuning; on code generation it reaches 57.2% on MBPP and 80.4% on HumanEval while full fine-tuning drops below the zero-shot baseline. We also demonstrate cross-architecture transfer on protein binding classification with ESM2-650M and cross-entropy training. BoostLoRA is, to our knowledge, the first PEFT method whose effective rank grows with training, separating per-round parameter cost from total representational capacity.
- [682] arXiv:2605.01140 (replaced) [pdf, html, other]
-
Title: SoCal: A Language for Memory-Layout Factorization of Recursive DatatypesVidush Singhal, Mikah Tal Kainen, Logan Anderson, Artem Pelenitsyn, Michael Vollmer, Michael H. Borkowski, Milind KulkarniSubjects: Programming Languages (cs.PL); Performance (cs.PF)
Array-of-structures (AoS) to structure-of-arrays (SoA) is a classic compiler transformation that improves memory locality and enables data-parallel execution. Existing AoS-to-SoA transformations target regular, array-based programs in imperative languages, such as C. However, many applications traverse tree-shaped data such as ASTs in compilers, DOM trees in browsers, and k-d trees in scientific workloads. Prior work speeds up functional programs over trees by serializing algebraic datatypes (ADTs) into one contiguous buffer. Such a representation interleaves the fields of every node, similar to an AoS layout. We introduce a factored layout of ADTs, which stores its fields in separate buffers, producing a layout that resembles SoA. In a factored layout, a traversal can focus on relevant fields and ignore the rest, similar to extracting a column from a SoA representation; in this setting, traversals over scalar fields can be vectorized. Working with raw byte arrays is error-prone, though: wrong indexing can silently read bytes from a wrong field or write past the end of a value. We therefore introduce SoCal, a type-safe language for expressing traversals over factored ADTs. Our compiler, Colobus, transforms a simple functional language with ADTs reminiscent of Haskell to C, using SoCal as the intermediate representation. Colobus generates traversals over byte arrays in the factored representation and performs optimizations such as field skipping and vectorization. Across a suite of benchmarks, Colobus shows the average speedups of 3.8x and 5.2x over fold and map passes, respectively.
- [683] arXiv:2605.04081 (replaced) [pdf, html, other]
-
Title: Time series causal discovery with variable lagsSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Causal Bayesian Networks (CBNs) are a powerful tool for reasoning under uncertainty about complex real-world problems. Such problems evolve over time, responding to external shocks as they occur. To support decision-making, CBNs require a cause-and-effect map of the variables under consideration, known as the network's structure. Learning the graphical structure of a causal model from data remains challenging; learning it from time-series data is even harder because dependencies may arise at different time lags. We construct a method that jointly learns a temporal ordering and a dominant lag for each selected ordered variable pair. Our contribution is a decomposable BIC-style objective score for mixed data that combines node-specific sample sizes with a penalty favouring recent lags. We optimise this objective using a Tabu search augmented with change-lag moves and greedy per-child lag tuning, while its node-wise decomposition supports cached and parallel neighbourhood evaluation. We establish acyclicity and complexity results and derive recovery limits when effects are distributed across multiple lags. In simulations, we show that the proposed method outperforms an inflexible lag-1 score-based learner in most evaluated settings, whereas most specialised methods achieve higher recovery accuracy on data that satisfy their assumptions. On a real-world UK COVID-19 dataset, the method learnt an expected short-lag-skewed structure; nevertheless, 46.8% of selected edges had lags greater than one, showing that longer delays can be retained when their improved fit offsets the penalty. Under its no-contemporaneous-edges and causal-sufficiency assumptions, the method provides a computationally tractable subroutine for edge-specific single-lag optimisation, complementing score-based causal discovery methods.
- [684] arXiv:2605.05043 (replaced) [pdf, html, other]
-
Title: Finding accurate eigenvalues and eigenvectors of positive semi-definite matrices given a subspaceComments: Revised based on reviewer feedbackSubjects: Numerical Analysis (math.NA)
We revisit a classical problem in numerical linear algebra: given a $k$-dimensional subspace $\mathcal{Q}$ with orthonormal basis $Q$ that approximates the leading eigenspace of an $n\times n$ positive semi-definite matrix $A$, the goal is to extract high-accuracy eigenvalues. The Rayleigh-Ritz (RR) method is the standard algorithm for the task, which has been shown to be optimal in several ways (when $A$ is symmetric, not necessarily positive semi-definite $A\succeq 0$). In this paper, we show that when $A\succeq 0$, alternative methods can outperform RR, including Nystr{ö}m and one-sided projected SVD, while having the same computational complexity; that is, the main cost is in computing $AQ$, plus an $O(nk^2)$ term. In particular, we advocate the use of Nystr{ö}m's method, quantitatively showing that the approximate leading eigenvalues have better accuracy than RR, and the improvement can be arbitrarily large. The significant accuracy improvement occurs for the first $i$ eigenvalues if $\lambda_i \gg \lambda_{k+1}$ and uniformly for all $k$ eigenvalues if $\lambda_k \gg \lambda_{k+1}$. This theoretical observation can be extended to the case where we extract $k$ leading eigenvalues from a larger-dimensional subspace. A similar improvement is numerically observed for the purpose of approximating the leading eigenvectors. In contrast, when the target eigenvalues are the trailing ones, the situation is reversed, and the Nystr{ö}m method performs poorly; we suggest a remedy based on a shift for this situation.
- [685] arXiv:2605.05623 (replaced) [pdf, html, other]
-
Title: Retrieval of Coastal Biogeochemical Parameters From Near-Surface Hyperspectral Remote Sensing Reflectance Using Physics-Aware Meta-LearningYiqing Guo, Nagur R. C. Cherukuru, Eric A. Lehmann, S. L. Kesav Unnithan, Tim J. Malthus, Gemma Kerrisk, Xiubin Qi, Faisal Islam, Tisham Dhar, Mark J. DoubellSubjects: Machine Learning (cs.LG)
Hyperspectral in situ sensing has shown promise in retrieving aquatic water quality (WQ) parameters, such as total suspended solids (TSS), dissolved organic carbon (DOC), and total chlorophyll-a (TChl-a), for cost-effective management of coastal WQ. However, generalising such retrieval algorithms across water bodies remains challenging, as the relationship between remote sensing reflectance ($R_{rs}$) and WQ parameters can vary considerably from one region to another due to regional distinctions in environmental conditions and biogeochemistry that lead to different WQ ranges and bio-optical properties. In this study, we propose a two-stage physics-aware meta-learning framework for retrieving coastal WQ parameters from near-surface $R_{rs}$ observations. Unlike purely data-driven approaches trained solely on paired observations, the proposed framework incorporates physical knowledge through a two-stage learning process. In the first stage, a region-agnostic base model is pretrained using meta-learning on synthetic WQ--$R_{rs}$ samples generated by a bio-optical forward model parameterised with an in situ bio-optical spectral library. In the second stage, the pretrained model is fine-tuned using local samples to adapt to regional conditions. To evaluate the proposed approach, we collected in situ hyperspectral $R_{rs}$ and WQ measurements from five geographically distinct sites in Australian coastal waters.
- [686] arXiv:2605.07398 (replaced) [pdf, html, other]
-
Title: Exposing and Mitigating Temporal Attack in Deepfake Video DetectionZheyuan Gu, Minghao Shao, Zhen Wang, Keyu Mao, Ailiang Lin, Mingyou Liang, Mingkun Xu, Shijie Zhang, Hao Jiang, Yusong WangSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
While spatiotemporal deepfake detectors achieve high AUC, our experiments reveal their susceptibility to evasion attacks. These models tend to overfit on fragile temporal spectrum cues, rather than learning robust semantic causality. To mitigate this vulnerability, we propose SpInShield, a temporal spectral-invariant defense framework explicitly designed to decouple semantic motion from manipulatable spectral artifacts. We propose a learnable spectral adversary that dynamically synthesizes severe spectral deformations, simulating extreme attack scenarios. By employing a shortcut suppression optimization strategy, SpInShield compels the encoder to extract reliable forensic cues while purging unstable spectral statistics from the latent space. Experiments show that SpInShield obtains competitive performance on widely used datasets and outperforms the strongest baseline by 21.30 percentage points in AUC under simulated amplitude spectral attacks.
- [687] arXiv:2605.07402 (replaced) [pdf, html, other]
-
Title: InsHuman: Towards Natural and Identity-Preserving Human InsertionSubjects: Computer Vision and Pattern Recognition (cs.CV)
Human insertion aims to naturally place specific individuals into a target background. Although existing image editing models may have such ability, they often produce failure cases, including inappropriate human pose in new background, inconsistent number of people, and modified facial identity. Moreover, publicly available human datasets often lack full-body portraits and realistic physical interaction between humans and their background. To address these challenges, we propose InsHuman for natural and identity-preserving human insertion. Specifically, we propose Human-Background Adaptive Fusion (HBAF), which detects foreground humans to obtain a binary mask and applies region-aware weighting to align the human regions between predicted and ground-truth latents, ensuring the person's pose, count, and overall appearance are coherently adapted to the target this http URL further propose Face-to-Face ID-Preserving (FFIP), which detects and matches faces between the generated image and the source image in terms of face recognition features to enforce identity consistency for each this http URL addition, we propose Bidirectional Data Pairing (BDP) strategy to construct BDP-InsHuman, a high-quality dataset with realistic human-background interactions. Experiments demonstrate that InsHuman achieves significant improvements in generating plausible images while keeping human identity unchanged.
- [688] arXiv:2605.07537 (replaced) [pdf, html, other]
-
Title: Multi-Environment POMDPs with Finite-Horizon ObjectivesComments: 33 pages, accepted at NeurIPS 2026 (Main Track)Subjects: Artificial Intelligence (cs.AI)
Partially Observable Markov Decision Processes (POMDPs) are systems in which one agent interacts with a stochastic environment, and receives only partial information about the current state. In a multi-environment POMDP (MEPOMDP), the initial state is unknown, and assumed to be adversarially chosen. In this work we focus on computing the optimal value and policy in MEPOMDPs with finite-horizon objectives. That problem is known to be PSPACE-complete in POMDPs. The extension is nontrivial: MEPOMDPs require reasoning about beliefs and payoffs across environments, and deterministic policies are no longer enough for optimal payoff. While the lower bound transfers immediately, the upper bound requires new techniques: we prove that mixtures of at most $n$ deterministic policies suffice for $n$ environments and show how to explore them without storing exponential-size policy descriptions. Our main results are as follows: (1) we establish that it is also PSPACE-complete in the more general setting of MEPOMDPs; (2) we present a practical algorithm and evaluate it on classical benchmarks, significantly outperforming the only previously known algorithm.
- [689] arXiv:2605.09028 (replaced) [pdf, html, other]
-
Title: Diagnosing and Mitigating Domain Shift in Permission-Based Android Malware DetectionSubjects: Machine Learning (cs.LG)
Machine learning-based Android malware detectors often fail in real-world deployment due to domain shift, where models trained on one data source perform poorly on applications from another. This paper presents a comprehensive study on the generalizability and interpretability of permission-based detectors under cross-domain conditions. Using two complementary datasets (PerMalDroid and NATICUSdroid) and five ensemble classifiers, we first establish an intra-domain baseline, where models achieve over 92% accuracy, and then quantify a severe asymmetric performance drop. While models trained on PerMalDroid generalize well to NATICUSdroid (86% accuracy), the reverse direction sees a drastic drop to 73% accuracy. Explainable AI analysis reveals bimodal feature distributions and shows that feature importance is highly unstable, with key permissions losing or gaining influence across domains. The predictive feature sets for different domains are fundamentally mismatched, as models rely on different, dataset-specific permissions. Most importantly, an ablation study demonstrates that for most models, training on a noisy feature set leads to poor generalization, confirming that domain-specific artifacts are a greater obstacle than missing features. To mitigate this, we validate a hybrid training strategy based on the intersection of common features and successfully recover cross-domain performance, achieving 88% accuracy on PerMalDroid and maintaining 97% on NATICUSdroid. These findings highlight the importance of explainable, cross-domain-robust malware detection systems and provide a practical pathway toward improving real-world deployment of permission-based Android malware detectors.
- [690] arXiv:2605.09159 (replaced) [pdf, html, other]
-
Title: Do LLMs Experience an Internal Polylogue? Investigating Reasoning through the Lens of PersonasSubjects: Artificial Intelligence (cs.AI)
Recent work shows that large language models (LLMs) encode behavioural traits ("personas") as linear directions in activation space, often called "persona vectors". Prior work has used such directions as static handles for behavioural steering. We instead treat them as dynamic signals: probes we can monitor and intervene on as reasoning unfolds. We use the term polylogue to denote the time series of alignments between persona vectors and hidden activations over the course of generation. Experiments across four open-weight models show that polylogue features predict correctness on MMLU-Pro on par with low-dimensional activation baselines, while remaining interpretable through their associated persona directions. They also suggest concrete steering targets, namely which latent directions to modulate at different stages of a response. We instantiate this as a simple paragraph-conditioned intervention that improves accuracy on three of four models but degrades the fourth, suggesting that stage-aware latent steering is possible but not yet robust. Together, this positions the polylogue as an interpretable tool for reasoning-time monitoring and intervention.
- [691] arXiv:2605.09549 (replaced) [pdf, html, other]
-
Title: When Adaptation Fails: A Gradient-Based Diagnosis of Collapsed Gating in Vision-Language Prompt LearningComments: Revised technical descriptions and diagnostic interpretations; clarified gate definitions, effective-length normalization, and regularization formulas. Reported benchmark results remain unchangedSubjects: Machine Learning (cs.LG)
Adaptive prompting mechanisms have been proposed to enhance vision-language models by dynamically tailoring prompts to inputs. However, in frozen few-shot prompt learning with CLIP-style backbones, we systematically observe that adaptive gates and prompt-selection modules often collapse: they produce nearly constant outputs, contribute negligible gradient signals, and frequently fail to outperform fixed prompts. To further explore this issue, we present a systematic diagnostic study to uncover the underlying causes and conditions of adaptation failure. Through controlled experiments across datasets and multiple prompt learning architectures, we identify two recurring failure modes: gradient magnitude imbalance and gate degradation. Our findings invite a re-examination of indiscriminately adding architectural complexity in parameter-efficient learning and clarify when prompt-level adaptive gating is, and is not, effective in this regime.
- [692] arXiv:2605.10005 (replaced) [pdf, html, other]
-
Title: Hybrid Mechanical and Agentic Specification InferenceSubjects: Programming Languages (cs.PL); Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO); Software Engineering (cs.SE)
Adoption of formal verification has long been hindered by the cost of manual specification. To prove high-level properties, one must also write contracts for supporting functions, many of which are mechanical and low-level. Specification inference automatically derives such contracts from the program. We present a hybrid of mechanical and agentic inference for Move, a smart contract language, integrated into its verifier, the Move Prover. A large language model (LLM) agent drives inference and verification through a tool that calls the Move Prover and derives specifications from code by weakest-precondition (WP) analysis. We evaluate our approach on 26 targets, 24 of them extracted from public Move functions of the Aptos framework that have no specifications, under a frontier model and a mid-tier model of the same family (Sol 5.6 and Terra 5.6). With and without WP, the agent completes over 90% of its runs, so agentic inference is effective on its own. Adding WP lowers its cost by 11-21% overall and by 40-64% on targets without loops.
- [693] arXiv:2605.11490 (replaced) [pdf, html, other]
-
Title: Adaptive Calibration in Non-Stationary EnvironmentsComments: 50 pages, NeurIPS 2026Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Making calibrated online predictions is a central challenge in modern AI systems. Much of the existing literature focuses on fully adversarial environments where outcomes may be arbitrary, leading to conservative algorithms that can perform suboptimally in more benign settings, such as when outcomes are nearly stationary. This gap raises a natural question: can we design online prediction algorithms whose calibration error automatically adapts to the degree of non-stationarity in the environment, smoothly interpolating between i.i.d. and adversarial regimes? We answer this question in the affirmative and develop a suite of algorithms that achieve adaptive calibration guarantees under multiple calibration measures. Specifically, with $T$ being the number of rounds, $W$ being the unknown number of i.i.d. segments of the environment, and $C\in[0,T]$ being another unknown non-stationary measure defined as the minimal $\ell_1$ deviation of the mean outcomes, our algorithms attain $\widetilde{\mathcal{O}}(\min\{\sqrt{T}+(TC)^{\frac{1}{3}}, \sqrt{WT}\})$ for $\ell_1$ calibration error and $\widetilde{\mathcal{O}}(\min\{(1+C)^{\frac{1}{3}}, W\})$ for both $\ell_2$ and pseudo KL calibration error. These bounds match the optimal rates in the stationary case ($C=0$ and $W=1$) and recover known guarantees in the fully adversarial regime ($C, W=\Omega(T)$). Our approach builds on and extends prior work (Hu et al., 2026; Luo et al., 2025), introducing an epoch-based scheduling together with a novel non-uniform partition of the prediction space that allocates finer resolution near the underlying ground truth.
- [694] arXiv:2605.12063 (replaced) [pdf, html, other]
-
Title: Memory Constrained Adversarial Hypothesis TestingSubjects: Information Theory (cs.IT)
We study adversarial binary hypothesis testing under memory constraints. The test is a time-invariant randomized finite state machine (FSM) with S states. Associated with each hypothesis is a set of distributions. Given the hypothesis, the distribution of each sample is chosen from the set associated with the hypothesis by an adversary who has access to past samples and the history of states of the FSM so far. We obtain upper and lower bounds on the minimax asymptotic error probability that are tight to within one state. These bounds determine the exact error exponent in S, and coincide under a weak least-favourability condition. For time-invariant deterministic FSMs, we also extend known achievable error-exponent bounds from the non-adversarial Bernoulli setting to the adversarial case with parameters in two separated intervals.
- [695] arXiv:2605.13276 (replaced) [pdf, html, other]
-
Title: D-VLA: A High-Concurrency Distributed Asynchronous Reinforcement Learning Framework for Vision-Language-Action ModelsYucheng Guo, Yongjian Guo, Zhong Guan, Wen Huang, Haodong Yue, Shuai Di, Junwu Xiong, Yicheng Gong, Xi XiaoSubjects: Artificial Intelligence (cs.AI); Robotics (cs.RO)
The rapid evolution of Embodied AI has enabled Vision-Language-Action (VLA) models to excel in multimodal perception and task execution. However, applying Reinforcement Learning (RL) to these massive models in large-scale distributed environments faces severe systemic bottlenecks, primarily due to the resource conflict between high-fidelity physical simulation and the intensive VRAM/bandwidth demands of deep learning. This conflict often leaves overall throughput constrained by execution-phase inefficiencies. To address these challenges, we propose D-VLA, a high-concurrency, low-latency distributed RL framework for large-scale embodied foundation models. D-VLA introduces "Plane Decoupling," physically isolating high-frequency training data from low-frequency weight control to eliminate interference between simulation and optimization. We further design a four-thread asynchronous "Swimlane" pipeline, enabling full parallel overlap of sampling, inference, gradient computation, and parameter distribution. Additionally, a dual-pool VRAM management model and topology-aware replication resolve memory fragmentation and optimize communication efficiency. Experiments on benchmarks like LIBERO show that D-VLA significantly outperforms mainstream RL frameworks in throughput and sampling efficiency for billion-parameter VLA models. In trillion-parameter scalability tests, our framework maintains exceptional stability and linear speedup, providing a robust system for high-performance general-purpose embodied agents.
- [696] arXiv:2605.13340 (replaced) [pdf, html, other]
-
Title: Shortcut Mitigation via Spurious-Positive SamplesComments: preprintSubjects: Machine Learning (cs.LG)
Shortcut mitigation strategies commonly rely on training data annotations, group-balanced held-out data or the presence of all groups, i.e., all combinations of (spurious) attributes and classes, in the training data. However, these requirements are rarely met in practice. We instead propose a method for targeted model analysis to identify a small set of instances in which the model relies on spurious attributes. Using that set and following ``this feature should not be used for prediction'' reasoning, we identify highly relevant neurons in an intermediate layer and regularize their impact. This ensures that models learn to depend on informative features rather than being right for the wrong reasons, thereby improving robustness without requiring additional balanced held-out data or annotations.
- [697] arXiv:2605.13348 (replaced) [pdf, html, other]
-
Title: Adequate Losses via Quantitative Linear LogicComments: Numerous tweaks and small corrections. To appear at POPL 2027Subjects: Logic in Computer Science (cs.LO); Logic (math.LO)
As neural components are increasingly embedded in existing symbolic software -- including safety-critical systems -- the question arises of how to specify and enforce the safety of the newly introduced neural parts. Unlike traditional logical specifications, these must be amenable not only to the standard Boolean interpretation, but also to training and optimisation. The latter calls for a quantitative interpretation of the logical syntax, subject to further requirements such as smoothness and differentiability. Moreover, the qualitative and quantitative sides of the logic must share a unifying proof-theoretic and categorical semantics. Finally, the new logic should link cleanly to the substructural and program logics that underpin the verification of existing symbolic programs. In this paper, we present a logic that ticks all of these boxes. We introduce a family of calculi, $p$QLL, indexed by a hardness degree $p$, prove a cut-elimination theorem for them, and establish completeness with respect to enriched residuated `soft' lattices. At $p = \infty$, $p$QLL reduces to multiplicative additive linear logic (MALL), and provability in $p$QLL converges to provability in MALL as $p \to \infty$. We express optimisation objectives in the syntax of this logic and prove the quantitative adequacy of neuro-symbolic loss functions -- a result that has eluded the neuro-symbolic machine learning community for nearly a decade.
- [698] arXiv:2605.16116 (replaced) [pdf, html, other]
-
Title: ShopGym: An Integrated Framework for Realistic Simulation and Scalable Benchmarking of E-Commerce Web AgentsChinmay Savadikar, Mingyu Zhao, Yuanzheng Zhu, Han Li, Shuang Xie, Alberto Castelo, Tianfu Wu, Lingyun WangComments: NeurIPS 2026 Evaluations and Datasets TrackSubjects: Artificial Intelligence (cs.AI)
Developing and evaluating e-commerce web agents requires environments that preserve meaningful task structure while enabling controllable, reproducible, and scalable scientific comparison. Existing methods force a tradeoff: live storefronts provide realism but are non-stationary, difficult to inspect, and irreproducible, while hand-built sandbox benchmarks provide control but cover only a narrow range of layouts, catalogs, policies, and interaction patterns. We argue that the core bottleneck is methodological: the field lacks a scalable way to construct evaluation settings that are simultaneously realistic, diverse, controllable, inspectable, and reproducible. We introduce ShopGym, an integrated framework for realistic simulation and scalable benchmarking of e-commerce web agents. Its simulation layer, ShopArena, converts live seed storefronts into self-contained sandbox shops through anonymized shop specifications and a staged, validated generation process. On top of these simulated storefronts, ShopGuru synthesizes benchmark tasks across seven skill categories, grounding each task in the shop's catalog, navigation structure, policies, and interaction affordances. Together, ShopArena and ShopGuru produce self-contained, resettable, inspectable, and stable evaluation artifacts that preserve structural properties and agent-evaluation signals relevant to shopping tasks. We validate the framework through graph-based structural analysis and agent-based behavioral evaluation with 224 generated tasks across six sandbox shops (three with synthetic data, three with real data). Results show that synthetic shops preserve key structural properties of live storefronts, and that agent success rates on synthetic twins closely track live storefronts, with matching per-skill profiles and failure modes. Generating ten shops shows the pipeline scales at about $60 per shop with declining human effort.
- [699] arXiv:2605.16573 (replaced) [pdf, html, other]
-
Title: Wavelet Flow Matching for Multi-Scale Physics EmulationGabriele Accarino, Juan Nathaniel, Carla Roesch, Pierre Gentine, Sara Shamekh, Duncan Watson-Parris, Viviana AcquavivaComments: Accepted at NeurIPS 2026; camera-ready version. Please check the project webpage for links to the source code, pre-trained weights, and datasets: this https URLSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Fluid Dynamics (physics.flu-dyn)
Accurate emulation of multi-scale physical systems governed by PDEs demands models that remain stable over long autoregressive rollouts while preserving fine-scale structures. Deterministic emulators produce overly smoothed predictions, while generative approaches better capture details but are costly. Latent-space generative models have emerged as a compromise but with the additional cost of separately pre-trained autoencoders. We propose Wavelet Flow Matching (WFM), a generative emulator that overcomes current trade-offs between cost and skill by performing optimal-transport flow matching directly in the multi-scale wavelet space. Rather than learning a latent compression, WFM leverages the hierarchical structure of a U-Net to jointly predict the transport velocities of a prescribed, lossless wavelet representation. On three challenging systems of chaotic fluid dynamics, WFM achieves superior spectral coherence and competitive or better long-horizon accuracy compared to state-of-the-art models. Our results position the wavelet space as an effective training-free representation for generative emulation of complex physical dynamics.
- [700] arXiv:2605.18020 (replaced) [pdf, html, other]
-
Title: Federated Learning by Utility-Constrained Stochastic Aggregation for Improving Rational ParticipationComments: Accepted to NeurIPS-2026Subjects: Machine Learning (cs.LG)
Federated Learning (FL) algorithms implicitly assume that clients passively comply with server-side orchestration, sharing local model updates upon the server's request. However, this overlooks an important aspect of real-world cross-silo environments: clients are often rational agents who may prioritize their own interests, such as local model performance, over that of the global model. In settings with significant statistical heterogeneity, rational clients may opt out of the federation if the perceived benefits of collaboration fail to meet their local utility thresholds. Such attrition degrades global model performance and can lead to the collapse of the federated training process. To avoid such a scenario, in this work, we introduce FedUCA (Federated Learning by Utility-Constrained Stochastic Aggregation for Improving Rational Participation), a non-monetary framework that establishes a non-payment channel through which clients communicate their willingness to participate via their utilities. The server acts as an optimizer, identifying aggregation parameters that satisfy clients' individual rationality (IR) constraints to sustain the global model performance. FedUCA achieves this through a stochastic aggregation mechanism over Dirichlet-sampled mixture strategies, which induces an implicit utility surplus that helps satisfy participation constraints. Supported by theory and extensive experiments, FedUCA's focus on participation feasibility improves client retention and global model performance.
- [701] arXiv:2605.21965 (replaced) [pdf, html, other]
-
Title: SpecHop: Continuous Speculation for Accelerating Multi-Hop Retrieval AgentsComments: Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)Subjects: Computation and Language (cs.CL)
Large language models increasingly use external tools such as web search and document retrieval to solve information-intensive tasks. However, multi-hop tool use in complex tasks introduces substantial latency, since the model must repeatedly wait for tool observations before continuing. We study how to accelerate such trajectories without changing the final trajectory the model would have taken without acceleration, assuming access to faster but less reliable speculator tools. We develop a theoretical framework for lossless speculation in multi-hop tool-use settings, characterizing the optimal achievable latency gain. We propose SpecHop, a continuous speculation framework that maintains multiple speculative threads, verifies predicted observations asynchronously as target tool outputs arrive, commits correct branches, and rolls back incorrect ones. This preserves accuracy while reducing wall-clock latency. We show that SpecHop can approach oracle latency gains with enough active threads. Empirically, on retrieval-augmented multi-hop tasks, SpecHop closely matches theoretical predictions and reduces latency by up to 40\% in some settings. Code: this https URL
- [702] arXiv:2605.22355 (replaced) [pdf, html, other]
-
Title: TransitLM: A Large-Scale Dataset and Benchmark for Map-Free Transit Route GenerationComments: Accepted at NeurIPS 2026 (Evaluations & Datasets Track)Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Public transit route planning traditionally depends on structured map infrastructure and complex routing engines, and no existing dataset supports training models to bypass this dependency. We present TransitLM, a large-scale dataset of nearly 13 million transit route planning records from four Chinese cities covering 120,845 stations and 13,666 lines, released as a continual pre-training corpus and benchmark data for three evaluation tasks with complementary metrics. Experiments show that an LLM trained on TransitLM produces structurally valid routes at high accuracy and implicitly grounds arbitrary GPS coordinates to appropriate stations without any explicit mapping. These results demonstrate that transit route planning can be learned entirely from data, enabling end-to-end, map-free route generation directly from origin-destination information. The dataset and benchmark are available at this https URL, with evaluation code at this https URL.
- [703] arXiv:2605.22432 (replaced) [pdf, html, other]
-
Title: AMUSE: Anytime Muon with Stable Gradient EvaluationComments: 55 pages, 28 figures. NeurIPS 2026 SpotlightSubjects: Machine Learning (cs.LG)
Modern deep learning commonly relies on AdamW with prescribed learning rate schedules, but recent works challenge both components: Schedule-Free optimization removes explicit schedules via iterate averaging, and Muon improves the update geometry by orthogonalizing momentum for matrix parameters. Despite Muon's strong empirical performance, its underlying mechanism remains partially understood. We study Muon through the river-valley loss landscape, where useful training progress occurs along a flat, low-curvature bulk subspace (the river), while high-curvature dominant directions form steep valley walls that induce oscillations. We empirically show that while Muon's orthogonalization accelerates river progress by increasing the bulk component, it also amplifies dominant-direction noise, causing oscillatory trajectories. Building on this, we propose Anytime MUon with Stable gradient Evaluation (AMUSE), which integrates Muon's rapid bulk progress with the stabilizing effect of Schedule-Free averaging. AMUSE uses a time-varying interpolation coefficient that initially evaluates gradients near the fast Muon sequence for rapid adaptation, then gradually shifts toward the stable averaged sequence to suppress valley-wall oscillations. As a result, AMUSE requires no learning rate schedules and supports anytime training. Across vision tasks and large language model pretraining, AMUSE consistently improves the performance-iteration Pareto frontier over (Schedule-Free) AdamW and Muon.
- [704] arXiv:2605.22781 (replaced) [pdf, html, other]
-
Title: DeltaBox: Millisecond-Scale Sandbox Checkpoint and Rollback for Stateful AI AgentsYunpeng Dong, Jingkai He, Shiqi Liu, Yuze Hou, Dong Du, Zhonghu Xu, Si Yu, Baochuan Yang, Yubin Xia, Haibo ChenComments: 16 pages, 9 figures, 3 tables. Accepted to SIGOPS ATC 2026Subjects: Operating Systems (cs.OS); Artificial Intelligence (cs.AI)
Emerging LLM-powered AI agent workloads, such as test-time tree search and reinforcement learning (RL), require high-frequency state exploration, relying on rapid checkpoint/restore (C/R) of the complete sandbox state, including files and process state. Existing mechanisms duplicate the entire state, causing hundreds of milliseconds to seconds of latency per C/R, which severely bottlenecks deep search and large-scale fan-outs. We observe that consecutive sandbox states between checkpoints differ only marginally. Therefore, instead of full duplication, a sandbox should only duplicate the changes between consecutive checkpoints (Key Insight), yet the lack of OS support makes this hard to realize. We propose a new OS-level abstraction, DeltaState, that delivers change-based C/R for agents through co-designed OS mechanisms. First, DeltaFS checkpoints file state by freezing the writable layer and inserting a new one, turning updates into copy-on-write and rollback into a layer switch. Second, DeltaCR quickly restores process state by forking from a frozen template process. A coordinator pairs the two so each checkpoint is one consistent (filesystem, memory) state. Building on these mechanisms, we present DeltaBox, an agent sandbox with millisecond-level C/R. On SWE-bench, DeltaBox hides its checkpoint work (10.83 ms) under inference and rolls back in a mean 1.86 ms, with the same warm-template fork driving low-cost RL rollout fan-out. DeltaBox is open-sourced at this https URL.
- [705] arXiv:2605.24366 (replaced) [pdf, html, other]
-
Title: SA-RAG: Structuring Noisy Conversations for Retrieval-Augmented GenerationKaiqiao Han, LuAn Tang, Renliang Sun, Peng Yuan, Wei Cheng, Haoyu Wang, Wei Wang, Yizhou Sun, Haifeng ChenSubjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Large Language Models (LLMs) have been widely adopted in conversational applications. However, their reliance on parametric knowledge limits reliability in real-world scenarios that require dynamic or domain-specific information. Retrieval-Augmented Generation (RAG) addresses this limitation by incorporating external knowledge during generation, but existing RAG methods often struggle with noisy or irrelevant contexts, as they typically assume that high-quality structures are already available. In this work, we propose Structure-aware Retrieval-Augmented Generation (SA-RAG), a principled structure induction and generation framework that constructs reliable intermediate structures directly from noisy conversational data. SA-RAG uses tables as compact and controllable structured representations. We introduce a quality-aware structure induction framework that jointly models semantic and structural consistency. Furthermore, we explore both training-free and training-based structure generation strategies to further improve structure quality while maintaining fidelity. Experiments on two noisy real-world datasets show that SA-RAG significantly outperforms existing RAG baselines. Further analysis highlights that high-quality structural representations are crucial for robust RAG applications.
- [706] arXiv:2605.25693 (replaced) [pdf, html, other]
-
Title: From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing AgentsRongsheng Zhang, Ruofan Hu, Weijie Chen, Jiji Tang, Junnan Ren, Wanying Wu, Xunuoyan Chen, Tangjie Lv, Tao Jin, Zhou ZhaoComments: Accepted by EMNLP 2026 FindingsSubjects: Computation and Language (cs.CL); Databases (cs.DB); Multiagent Systems (cs.MA)
While role-playing agents excel in short-term interactions, long-term conversations overwhelm context windows, motivating external memory frameworks. Current systems typically rely on persona-agnostic summarization, which records facts without persona-specific interpretation, yielding generic responses that compromise persona fidelity. To bridge this gap, we introduce RoleMemo, a dataset featuring four reasoning tasks where the factual fragments must be interpreted through the persona to reach the correct answer. Evaluation on RoleMemo exposes critical limitations of persona-agnostic frameworks. We thus propose DualMem, which decouples memory into two streams: factual cognition and persona-conditioned insight. Trained through Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL), our framework with a 4B-parameter model outperforms zero-shot persona-agnostic frameworks powered by DeepSeek-V3.2 for sustained persona fidelity. Our resources are available at this https URL.
- [707] arXiv:2605.29788 (replaced) [pdf, html, other]
-
Title: Certified Policy Optimisation for Nested Causal Bandits via PAC-Bayes RiskComments: Accepted at NeurIPS 2026 (Main Track)Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Critical sequential decisions are rarely single-timescale: a strategic decision causally shapes the context in which every subsequent tactical choice is made; standard bandit and reinforcement-learning theory does not capture this causal coupling between timescales. We formalise the problem class as Nested Contextual Causal Bandits (NCCBs), a hierarchical SCM where each level's action sets the next level's context distribution, and propose Nested Causal Thompson Sampling (NCTS), which draws one mechanism-factorised belief per episode and acts recursively under it. Our main theoretical result is a causal PAC-Bayesian excess-risk bound that certifies any candidate deployment policy from historic data alone, off-policy and anytime, answering the deployment question: can we trust this agent here, and at what risk? Experiments on a hierarchical SCM show that, against a matched RFF-GP joint regression on the same function class, the factorised SCM-mechanism posterior transfers significantly better zero-shot under exogenous distribution shifts, the recursive meta-to-inner commit significantly dominates the joint-commit alternative in distribution, and the certificate significantly contracts as offline data accumulates. Combining these results, we establish progressive certified handover, a safe-deployment method: each timescale flips from a legacy controller to NCTS when gains can be certified, independently of the others.
- [708] arXiv:2605.30564 (replaced) [pdf, html, other]
-
Title: Listing Even Cycles Faster than the Submodular-Width BarrierComments: 19 pages. Accepted to PODS 2027. To appear in Proceedings of the ACM on Management of Data (PACMMOD), Vol. 4, No. 5 (PODS), Article 305, November 2026Subjects: Databases (cs.DB); Data Structures and Algorithms (cs.DS)
A classic result of Alon, Yuster, and Zwick (AYZ, Algorithmica 1997) shows that all $2k$-cycles in an $m$-edge graph can be listed in $\tilde O(m^{2-1/k}+t)$ time, where $t$ is the output size. This bound underlies the {\em submodular width} of Marx (JACM 2013) and the PANDA framework of Abo Khamis, Ngo, and Suciu (PODS 2017), which extend AYZ to arbitrary conjunctive queries with degree constraints. A central open question is whether combinatorial algorithms can beat the submodular-width barrier.
Bringmann and Gorbachev (STOC 2025) gave lower-bound evidence that submodular width may be optimal for general conjunctive queries under combinatorial algorithms. The picture changes for $2k$-cycles on undirected graphs, whose queries have self-joins and symmetric EDBs: recent works improve on AYZ for even-cycle detection and listing. Pinning down the complexity of $C_{2k}$-detection and listing is thus a natural step toward overcoming the submodular-width barrier for such queries.
For detection, Dahlgaard, Knudsen, and St{ö}ckel (STOC 2017) solved $C_{2k}$-detection in $\tilde O(m^{2k/(k+1)})$ time. Listing is harder. Jin and Xu (STOC 2023), and independently Abboud, Khoury, Leibowitz, and Safier (FSTTCS 2023), listed 4-cycles in $\tilde O(m^{4/3}+t)$ time; Vassilevska~Williams and Westover (ITCS 2025) listed 6-cycles in $\tilde O(m^{8/5}+t)$ time, improving the AYZ bounds of $\tilde O(m^{3/2})$ and $\tilde O(m^{5/3})$. The general case has remained open for 30 years.
Building on these works, we list $2k$-cycles in $\tilde O(m^{(2k^2-k+1)/(k^2+1)}+t)$ time, improving AYZ for every $k\geq 3$. The key ingredient is an \emph{asymmetric supersaturation} result for even cycles. Our algorithms use only join and project operators over multiple tree-decomposition plans, making them naturally implementable in database systems, in contrast to prior BFS-based graph approaches. - [709] arXiv:2606.05373 (replaced) [pdf, html, other]
-
Title: Evidence-Guided Neural Architecture Selection under Uncertainty for Subject-Specific Blood Glucose ForecastingSubjects: Machine Learning (cs.LG); Biological Physics (physics.bio-ph)
Reliable neural architecture selection is an open challenge in time-series forecasting under limited, noisy, and heterogeneous data, where standard heuristic architecture design and validation approaches fail to ensure accurate and reliable prediction and generalization. We propose EVIDENT (EVidence-based IDEntification of Neural archiTectures), a framework for architecture selection that integrates Bayesian training, evidence-based ranking, and task-specific validation under uncertainty. The framework explores the candidate architecture pool and identifies the lowest-capacity model that satisfies a prescribed validation criterion. We demonstrate this method using temporal convolutional networks (TCNs) for individualized blood glucose forecasting in type 1 diabetes patients. The results show that EVIDENT systematically rejects both under- and over-parameterized TCN architectures on population-level diabetes data, while identifying models that generalize reliably to unseen patients. When multiple architectures are competitive, the framework further supports plausibility-weighted ensemble predictions that enhance predictive performance. Compared with a random-search baseline, EVIDENT identified smaller architectures with more consistent forecasting performance on unseen patients. These findings establish EVIDENT as a strategy to neural architecture discovery, enabling reliable model selection for high-consequence forecasting in data-limited and heterogeneous settings.
- [710] arXiv:2606.07239 (replaced) [pdf, html, other]
-
Title: Generative Molecular Morphing for Flexible-Size Design via Unbalanced Optimal TransportSubjects: Machine Learning (cs.LG)
The success of generative molecular design depends on a model's steerability toward high-reward samples. Because many molecular properties are intrinsically linked to molecular size, accurately capturing the joint distribution of properties and the number of atoms is essential. However, current diffusion and flow-based models fix the number of atoms, which ultimately limits their ability to navigate this complex relationship. To address this, we introduce Morph, a flexible-size generative model for conditional and unconditional 3D molecular design based on geometric graphs. By dynamically adapting size, Morph can seamlessly integrate existing structural priors, like scaffolds, and significantly enhances property steering. We show that Morph matches current fixed-size state-of-the-art models while significantly increasing sampling flexibility. We demonstrate out-of-distribution generation in regimes where previous models fail, paving the way for enhanced generative modeling for molecular design.
- [711] arXiv:2606.07923 (replaced) [pdf, html, other]
-
Title: Larch: Learned Query Optimization for Semantic PredicatesFuheng Zhao, Pawel Liskowski, Zihan Li, Benjamin Han, Puxuan Yu, Varich Boonsanong, Dimitris Tsirogiannis, Anupam DattaSubjects: Databases (cs.DB); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
With the advent of Large Language Models (LLMs), many database systems introduced semantic operators that enabled analytical queries over unstructured data (e.g. text, images, videos). Semantic operators typically incur high inference costs and latencies making semantic (AI) SQL queries challenging to apply on large scale datasets. At the same time, their semantic nature leads database engines to treat them as black boxes, making AISQL queries difficult to optimize. In this paper, we introduce Larch, a framework for optimizing the execution of semantic filters in AI SQL queries. Larch was inspired by two key observations: i) the high latency of semantic operators leaves significant room for computationally-heavy runtime optimization techniques, ii) unstructured data are typically accompanied by semantic information in the form of embeddings allowing for efficient semantic comparisons between AI_FILTER prompts and data values. Based on these two key observations, we present two Larch variants: Larch-A2C and Larch-Sel. Larch-A2C encodes arbitrary semantic filters expression tree using an embedding-augmented Gated Graph Neural Network and formulates the filter evaluation order as a Markov decision process. In contrast, Larch-Sel leverages a supervised learning model to predict filter selectivities, subsequently applying dynamic programming to find a near-optimal evaluation order for each input row. Evaluated across diverse real-world datasets and comprehensive synthetic workloads, both Larch variants always outperform existing semantic filter optimization techniques in terms of token usage on every dataset and workload. Our results demonstrate that Larch is robust across diverse workloads across text and image modalities, 6%--30% lower total token costs and 3x--19x lower token costs overhead than Palimpzest and Quest.
- [712] arXiv:2606.10944 (replaced) [pdf, html, other]
-
Title: Express Language ModelingSubjects: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS); Statistics Theory (math.ST); Methodology (stat.ME); Machine Learning (stat.ML)
We introduce a new tool, Express, for converting a non-causal attention approximation into a causal approximation with matching approximation guarantees. When combined with the state-of-the-art Thinformer approximation, Express improves upon the best known causal attention guarantees, delivering $\log^{3/2}(n)/s$ approximation error with only $O(s)$ memory and $O(s^2 \log^2(n))$ compression overhead for a sequence of length $n$. We pair these developments with an efficient I/O-aware Triton implementation, demonstrate substantial speedups over FlashAttention 2, and use Express to overcome four resource bottlenecks in the language modeling pipeline: long-context prefill, KV cache compression, long-form memory-constrained decoding, and long-form compute-constrained decoding.
- [713] arXiv:2606.13880 (replaced) [pdf, html, other]
-
Title: A Horizon-Specific Attribute-Conditioned Neural Network for Estimating Health-State Transition Probabilities in Irregular Longitudinal Data: The LANTERN FrameworkComments: 35 pages, 8 figuresSubjects: Machine Learning (cs.LG); Risk Management (q-fin.RM)
Accurate estimation of health-state transition probabilities is essential for long-term care insurance modeling. Conventional actuarial multi-state models provide an interpretable framework but may be restrictive in capturing nonlinear relationships and accumulated health histories from irregular longitudinal observations. This paper proposes LANTERN, a history-dependent neural framework that integrates recurrent health representations, elapsed observation time, and adaptive demographic and socioeconomic attribute conditioning to estimate horizon-specific transition probabilities among Healthy, Mild disability, Severe disability, and Death states. Using longitudinal Health and Retirement Study data, we evaluate our model against a recurrent neural network, multinomial logistic regression, gradient-boosted trees, and an actuarial continuous-time Markov chain across one-, two-, and three-year forecast horizons. At the primary two-year horizon, LANTERN achieves competitive predictive accuracy and calibration, with improvements over selected benchmarks. Its aggregated transition-matrix estimates outperform multinomial logistic regression and are comparable to those of the strongest remaining benchmarks. These findings suggest that history-conditioned neural estimation provides a flexible and competitive approach to estimating horizon-specific health-state transition probabilities, with potential applications in actuarial long-term care modeling.
- [714] arXiv:2606.16119 (replaced) [pdf, html, other]
-
Title: EdgeZSAD: Value-Range Quantization for Deploying Zero-Shot Anomaly Detection on Edge Neural Processing UnitsComments: 22 pages. Substantially revised version with a focus on Value-Range Quantization and native edge deployment evaluationSubjects: Computer Vision and Pattern Recognition (cs.CV)
Quantizing zero-shot industrial anomaly detectors for edge neural processing units can degrade defect detection and localization. We present EdgeZSAD (Edge Zero-Shot Anomaly Detection), a deployment framework whose methodological contribution is Value-Range Quantization (VRQ). We diagnose how attention Value tensors inherit overly wide ranges from fused query-key-value projections and assign independent ranges calibrated on source data, without retraining or changing quantized weights. The engineering application is native deployment of a foundation-model detector for industrial inspection, evaluated on three public benchmarks and Qualcomm edge devices. Under minimum-maximum and mean-squared-error calibration, VRQ improves image-level discrimination and per-region overlap. In the primary calibration set, it recovers 90-98% of the loss in combined image and pixel discrimination relative to half precision, with a small increase in latency over baseline quantization. Matched tensor overrides reproduce the recovery without adding graph nodes, identifying independent Value-range assignment as the effective intervention in the evaluated runtime family.
- [715] arXiv:2606.16490 (replaced) [pdf, html, other]
-
Title: Robots that Collaborate: Sequential Asymmetric Imitation for Learning Coupled Robot PoliciesSubjects: Robotics (cs.RO)
Collaborative mobile manipulation requires robots to coordinate with a partially observed partner while physically interacting through shared objects. This is difficult because failures often arise not from poor local skills, but from mistimed waiting, yielding, pulling, releasing, or repositioning. We study this problem with two bimanual mobile manipulators coupled through rigid and deformable objects. We propose Sequential Asymmetric Imitation (SAI), a single-teleoperator curriculum for learning coupled multi-robot behaviors without synchronized dual-operator demonstrations or explicit inter-robot communication. SAI trains Robot A from unilateral demonstrations with a compliant human partner, trains Robot B against the deployed Robot A policy, and then refines Robot A using sparse interventions near coordination failures. This staged process exposes the policies to increasingly realistic partner behaviors, including delay, phase mismatch,insufficient yielding, and interaction conflict. Across real-world dual-robot manipulation tasks, SAI improves task success, phase synchronization, and partner-contingent yielding over independent imitation and curriculum-ablation baselines. These results suggest that physically coupled collaboration can be learned through the structure of the imitation curriculum, rather than through synchronized multi-operator demonstrations or explicit coordination mechanisms. More videos on project page:this http URL
- [716] arXiv:2606.16613 (replaced) [pdf, html, other]
-
Title: CoffeeBench: Benchmarking Long-Horizon LLM Agents in Heterogeneous Multi-Agent EconomiesIssa Sugiura, Daichi Hattori, Kazuo Araragi, Keita Ogawa, Shota Onose, Taro Makino, Teppei Usuki, Takashi IshidaComments: Accepted to NeurIPS 2026 (Evaluations and Datasets Track)Subjects: Artificial Intelligence (cs.AI)
LLM agents are increasingly capable of handling complex, long-horizon tasks and may eventually take on business management, a role traditionally performed by humans. Deploying them safely requires monitoring their capabilities and behavior in environments that reflect real economies. We introduce CoffeeBench, a benchmark built on a compact simulation of a coffee supply chain with realistic business mechanics. Six autonomous LLM coding agents run its farmers, roasters, and retailers, and negotiate and trade with one another to maximize their own net income over 90 simulated days. We evaluate eight recent proprietary and open-weight models by letting each run one roaster while a fixed background model runs the other five firms, and scoring that roaster's net income. Claude Sonnet 5.5 earns the highest net income, closely followed by GPT-6 Astra. The leaders sell at higher prices, and the best model also buys its inputs far more cheaply, whereas weaker models trade at thin margins. Since performance targets are a common cause of accounting fraud in human firms, we also test whether agents under KPI pressure inflate their numbers. Under a revenue target, circular trades occur for some pairs of models: one agent proposes that the two rival firms sell goods back and forth at inflated prices so that both book revenue, and the rival accepts. We hope CoffeeBench serves as an early step toward a future in which LLM agents manage firms without misconduct. We release our code and agent trajectories.
- [717] arXiv:2606.21387 (replaced) [pdf, html, other]
-
Title: Long-Distance Real-World Navigation of the Legged-Wheeled Robot Go2-W Using Deep Reinforcement LearningSubjects: Robotics (cs.RO)
Legged-wheeled robots have been extensively studied for their potential to combine the efficient flat-ground mobility of wheels with the rough-terrain adaptability of legs. However, their application to long-distance autonomous navigation in real environments remains limited. This paper reports our efforts to build a deep reinforcement learning-based locomotion controller and an autonomous navigation system for the commercially available Go2-W legged-wheeled robot. The study focuses on the integrated system and the practical challenges associated with its sustained operation over long distances in real-world environments. For locomotion control, we extended our proprioception-only policy, previously developed for quadruped robots, to Go2-W, a 16-degree-of-freedom legged-wheeled robot. We additionally found that wheeled locomotion concentrates the load on the hip joints, resulting in heat concentration that hinders sustained travel. To address this issue, we developed a policy with periodic foot-lifting motion that distributes the joint loads and mitigates heat concentration. We evaluated the integrated system at the Tsukuba Challenge 2025. The robot autonomously traversed an approximately \SI{2.8}{\kilo\meter} route that included sidewalks, a park, and stairs without stopping owing to overheating.
- [718] arXiv:2606.24537 (replaced) [pdf, html, other]
-
Title: Importance Sampling for Event Discovery via GuessworkComments: Short version submitted to ITW 2026. This is a longer version after being accepted to ITW, correcting reviewers comments, organizing and rephrasing, and adding more complete proofs, comments, etcSubjects: Information Theory (cs.IT)
Traditional importance sampling (IS) is designed to estimate rare-event probabilities by minimizing estimator variance. However, many applications prioritize rapid discovery: the generation of a trajectory within a rare set $A_n$. This requires a shift from ensemble-based estimation to a design principle focused on the hitting time $\tau_{A_n} := \inf\{t \ge 1 : Y_t^n \in A_n\}$.
We formalize a Quality of Discovery problem as the problem of minimizing the description length (surprisal) of the discovered trajectory under the nominal model $p$. We prove that minimizing this description length is equivalent to minimizing the nominal rank exponent $J_{\mathrm{rank}}(q_n) := \lim_{n\to\infty} \frac{1}{n} \log G_n(Y^n)$, where $G_n(x^n)$ is the guesswork of sequence $x^n$. For i.i.d.\ models and type-defined rare sets $\Gamma$, we show that while classical IS targets the mass-dominating type $Q_{\mathrm{IS}}^* \in \arg\min_{Q \in \Gamma} D(Q\|p)$, discovery optimality is achieved by $Q_{\mathrm{GW}}^* \in \arg\min_{Q \in \Gamma} [H(Q) + D(Q\|p)]$. This framework identifies a fundamental rule: minimizing the guesswork exponent ensures the discovered sequence is the "least surprising" representative of the set relative to the nominal model's search order. We further demonstrate that under budgetary constraints, this exponent serves as a lexicographic tie-breaker when the hitting-time minimizer is not unique. This establishes $H(Q) + D(Q\|p)$ as a natural objective for discovery-based importance sampling, providing a formal bridge between randomized sampling and systematic search. - [719] arXiv:2606.28109 (replaced) [pdf, html, other]
-
Title: MMAO: A Metabolic Multi-Agent Optimizer with Endogenous Resource Allocation for Continuous and Discrete OptimizationComments: 12 pagesSubjects: Neural and Evolutionary Computing (cs.NE); Multiagent Systems (cs.MA)
Traditional meta-heuristics often rely on fixed population sizes, manually chosen search scales, and externally attached parameter-control modules. This paper presents the Metabolic Multi-Agent Optimizer (MMAO), a cross-domain optimization framework in which adaptation is derived endogenously from a private-public metabolic resource loop. Each agent carries internal energy, a continuous role state, motion or structural memory, and local search history, while the population shares a communal resource pool. Fitness improvements are converted into normalized metabolic gains through a robust progress scale and a recent success statistic; the same closed loop then regulates sensing intensity, search amplitude, role drift, branching, pruning, respawning, and elite reinvestment. In the continuous setting, MMAO uses energy-regulated symmetric zero-order probing and role-interpolated motion. In the discrete setting, the same control law is instantiated through structural sensing, local route improvement, guided perturbation, and energy-weighted edge reuse. The paper combines an implementation-faithful formulation with a reproducible experimental study on a CEC2017 subset (10D/30D, 20 seeds) and five TSPLIB instances (100 discrete runs in total). The current evidence supports MMAO primarily as a parameter-light, self-calibrating optimization framework whose main validated originality lies in metabolically endogenous resource allocation across heterogeneous search behaviors, rather than as a universally superior optimizer.
- [720] arXiv:2606.28228 (replaced) [pdf, html, other]
-
Title: Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion ShiftsComments: Accepted at NeurIPS 2026 (camera-ready version). 53 pages, 15 figuresSubjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Causal representation learning for time series has developed strong identifiability results in discrete-time latent causal models, but identifiability in continuous-time latent stochastic differential equation (SDE) models remains largely open. We address this gap using environment-induced shifts in diffusion covariance. We study additive-noise latent SDEs observed through an unknown nonlinear diffeomorphism, with shared drift but environment-specific diffusion covariance. We show that two diagonal diffusion regimes with pairwise distinct coordinate-wise variance ratios identify the latent coordinates up to permutation, coordinate-wise scaling, and a possible constant shift, without any sparsity assumption on the drift. We first prove this result for linear Ornstein-Uhlenbeck systems and then extend it to general additive-noise latent SDEs. Under mild smoothness, the instantaneous drift-Jacobian causal graph is identifiable up to the same permutation. We propose a two-stage estimator for latent disentanglement and optional graph recovery; experiments on synthetic systems confirm the predicted identifiability boundary, and an application to Hardanger Bridge monitoring data illustrates the approach on real sensor trajectories.
- [721] arXiv:2606.29142 (replaced) [pdf, html, other]
-
Title: Agent Security Meets Regulatory Reality -- A Practitioner Systematization of Autonomous-Agent Threats and Controls in Regulated Financial SystemsComments: Substantially revised and expanded version. Added five reference algorithms, a formal execution model, theorem and proofs, security and performance analysis, and an open-source reference implementation. Expanded authorship to reflect substantial contributions from two additional coauthorsSubjects: Computers and Society (cs.CY); Software Engineering (cs.SE)
Large language model agents are entering regulated financial systems, yet research on their attack surface remains predominantly laboratory-based, and practitioner guidance is neither peer-reviewed nor connected to a formal threat model. Using production evidence, we map six agentic threat categories, namely prompt injection, identity and authorization, action auditability, tool abuse, data residency, and boundary policy enforcement, to the specific control obligations imposed by US and EU financial regulation (ECOA and Regulation B, EU AI Act, GDPR Article 22, and FINRA's 2026 agent guidance), showing how legal accountability increases their operational significance. We distill these obligations into five runtime control invariants with reference algorithms that enforce them at auditable architectural boundaries. We develop a formal model expressing each invariant as a predicate over a case-bound event chain and prove the audit reconstruction gate is sound and complete with respect to these invariants: compliance is decidable from the case identifier alone. We show that two of the invariants constrain audit cost and throughput rather than constituting pure operational overhead. We document four architectural patterns from a production Know Your Customer (KYC) deployment for a consumer credit product that reduced multi-day manual processing to same-day automated resolution for approximately four in five cases. Finally, we report three negative results, two identified through internal audit or production observation rather than adversarial testing. We conclude that securing agents in regulated environments depends less on novel attack classes than on rendering auditability, least-privilege authorization, and boundary policy enforcement, with verifiable compliance and bounded audit cost at scale.
- [722] arXiv:2606.30237 (replaced) [pdf, html, other]
-
Title: Comparing Human and Automatic Recognition of Dutch Dysarthric Continuous Speech: A Case StudySubjects: Computation and Language (cs.CL)
In our goal to develop personalised dysarthric speech recognition (DSR) models, this study compared the recognition performances of human listeners and those of three state-of-the-art, off-the-shelf ASR systems (Whisper-large-V3, Google Chirp 3, and Omnilingual) on the recognition of Dutch continuous read and spontaneous speech from a single speaker with severe dysarthria. Results showed that both humans listeners and the three off-the-shelf ASR systems exhibit word error rates (WER) exceeding 70% on average, indicating that DSR is highly challenging for both humans and ASR systems. Fine-tuning on the dysarthric speech significantly reduced WER. Although overall WERs are still quite high (>23%), the personalised DSR models outperformed the human listeners, and performance is getting closer to being useful for supporting day-to-day communication of dysarthric speakers. Future research should focus on improving personalized DSR on spontaneous speech and longer utterances in the case of read speech, with a specific focus on particular phonemes.
- [723] arXiv:2606.30450 (replaced) [pdf, html, other]
-
Title: Minimal MMAO: A Resource-Closed-Loop Framework for Adaptive Metaheuristic SearchComments: 6 pagesSubjects: Neural and Evolutionary Computing (cs.NE); Multiagent Systems (cs.MA)
This paper presents the Metabolic Multi-Agent Optimizer (MMAO) as an adaptive metaheuristic built around endogenous resource circulation. The central premise is that search intensity, exploration--exploitation balance, and lifecycle turnover should be induced by a shared metabolic controller rather than by separately attached schedules. We formulate MMAO through bounded private energy, a communal budget, normalized reward, continuous role adaptation, and resource-financed branching and pruning. The method is then instantiated in both continuous and discrete domains and evaluated on a matched small-scale suite including Sphere, Rastrigin, a synthetic Euclidean TSP, and two TSPLIB instances. The results show a consistent pattern: the same metabolic loop remains workable across domains, the discrete realization remains relatively stable under a compact design, and continuous refinement quality is the main cost of keeping the method lean. These results support MMAO's framework formulation under simplification, with some loss of late-stage precision on harder continuous tasks.
- [724] arXiv:2606.31584 (replaced) [pdf, html, other]
-
Title: A Large-Scale Empirical Evaluation of MMAO Under Fair-Budget Continuous and Discrete BenchmarksComments: 9 pagesSubjects: Neural and Evolutionary Computing (cs.NE); Multiagent Systems (cs.MA)
This paper evaluates the Metabolic Multi-Agent Optimizer (MMAO) under a stricter empirical protocol rather than reintroducing the framework itself. The study asks whether MMAO's closed-loop resource-allocation principle remains credible under broader, more standard, and more explicitly budget-controlled continuous and discrete benchmarks. The main completed matrix covers eight CEC2017 functions at 10D and 30D with 20 seeds each, and five TSPLIB instances with 20 seeds each, together with stronger reproducible baselines including PSO-lite, ES-lite, and an iterated-greedy 2-opt route baseline. We further add trajectory-level diagnostics for communal budget, success rate, role evolution, and population turnover, plus an auxiliary OR-Library multiple-knapsack slice to extend the discrete evidence beyond routing. Under this protocol, MMAO clearly outperforms the external baseline set on the continuous side and on the TSPLIB side, while the ablation variants remain much closer to the full method than the external baselines are. The results support endogenous resource redistribution across the tested domains, while sharper mechanism isolation and broader competition-grade comparison remain open.
- [725] arXiv:2606.32034 (replaced) [pdf, html, other]
-
Title: QVal: Evaluating Dense Supervision Signals for Long-Horizon LLM AgentsSergio Hernández-Gutiérrez, Matteo Merler, Ilze Amanda Auzina, Joschka Strüber, Ameya Prabhu, Matthias BethgeComments: 10 pages, 6 figures in main text; 67 pages, 13 figures with appendixSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
LLM agents increasingly act over long horizons, where outcome-only rewards say little about which intermediate actions were good. Dense supervision methods address this by scoring intermediate steps, but they are usually evaluated through the downstream performance of a full training pipeline. This is expensive, confounds signal quality with other training choices, and means that methods are rarely compared on common ground. We introduce QVal, a training-free testbed that measures the $Q$-alignment of a dense signal: whether it orders the candidate actions in each state the same way as their returns under a reference policy do. We benchmark 21 methods from seven families with six open-weight backbones in four environments. Simple prompting baselines are the most $Q$-aligned on average, and performance clusters strongly by family. Code-generated signals do well only in small state and action spaces; embedding-based signals are weak; self-distillation signals are anti-aligned; and no method is $Q$-aligned in open-ended environments. When five of these signals guide an actor at test time by choosing among four proposed actions, $Q$-alignment strongly correlates with the gain in task success. QVal is designed to be easily extensible to new environments and methods, enabling researchers to compare new signals cheaply before committing to a training run.
- [726] arXiv:2607.00833 (replaced) [pdf, html, other]
-
Title: GenTUS: Table Union Search via Constrained Generation of Semantic Table IdentifiersSubjects: Databases (cs.DB)
Modern data lakes contain heterogeneous tables, and the information needed for a task is often scattered across many tables with different schemas, sources, and naming conventions. Table union search (TUS) retrieves tables whose rows are semantically compatible with a query table for union, supporting data discovery, enrichment, and downstream analytics. Existing TUS methods follow an encode-search-refine pipeline: they embed tables or columns, retrieve a candidate pool from a vector index, and refine it by query-candidate matching or reranking. As a result, a unionable table missed by candidate retrieval cannot be recovered later, and searching and refining a pool for every query adds latency and storage costs that grow with the data lake. We propose GenTUS, a generative retrieval framework that reformulates TUS as constrained generation over discrete semantic table identifiers. GenTUS assigns each data-lake table a compact identifier learned under pairwise unionability supervision, which encourages unionable tables to share identifier prefixes, and trains a generator to produce the identifiers of unionable tables directly from the query table. At query time, prefix-constrained beam search guarantees that every generated identifier maps to an existing table, and its output order serves directly as the ranking, with no candidate pool to search or rerank. Experiments on seven public TUS benchmarks show that GenTUS achieves the best overall retrieval quality, with the largest gains on the larger and more diverse benchmarks, while requiring substantially lower online retrieval time and retrieval-artifact storage than prior TUS methods.
- [727] arXiv:2607.00846 (replaced) [pdf, html, other]
-
Title: MMAO-Dyn: A Metabolic Multi-Agent Optimizer for Dynamic OptimizationComments: 8 pagesSubjects: Neural and Evolutionary Computing (cs.NE)
This paper studies whether the Metabolic Multi-Agent Optimizer (MMAO) can be credibly derived into a dynamic-optimization method without replacing its core metabolic control loop by external adaptation modules. The proposed MMAO-Dyn maps private energy, communal budget, role drift, success feedback, and lifecycle turnover to a nonstationary setting in which environmental changes repeatedly invalidate previously useful local structure. We evaluate MMAO-Dyn on an 18-scenario synthetic dynamic continuous benchmark matrix covering shifted sphere, shifted Ackley, and shifted Rastrigin landscapes at $10D$, $20D$, and $30D$, with two change severities and 12 seeds per scenario. The comparison layer includes a generic MMAO variant without dynamic derivation, dynamic random search, dynamic PSO-lite, dynamic DE-lite, and three endogenous ablations. Across the full 216-run matrix, MMAO-Dyn attains mean offline error $28.07$, improving over Generic-MMAO ($29.36$), Dynamic-PSO-lite ($34.65$), Dynamic-DE-lite ($67.09$), and Dynamic-RandomSearch ($111.37$). The gains are clearest in aggregate robustness on sphere and Rastrigin families and in 10-step post-change recovery relative to the generic backbone, whereas the seed-aligned comparison with Dynamic-PSO-lite remains unfavorable in win-loss count and the NoMemoryRefresh ablation stays very close to the full method. The results support extending the metabolic loop to dynamic optimization, particularly through recovery-oriented resource redistribution.
- [728] arXiv:2607.01159 (replaced) [pdf, html, other]
-
Title: Online Fair Division Meets Reordering BuffersComments: A conference version of this work appears in the Proceedings of the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)Subjects: Computer Science and Game Theory (cs.GT); Discrete Mathematics (cs.DM); Data Structures and Algorithms (cs.DS)
We study the online fair division of indivisible mixed manna among agents with additive valuation functions. Under the standard online model, at each time step an indivisible item arrives; each agent may assign it a positive, negative, or zero value, and it must be irrevocably allocated, before the arrival of the next item. At the same time, we also wish to maintain some fairness guarantee, and in this work we focus on envy-freeness (EF) and one of its most prominent relaxations, envy-freeness up to one item (EF1). Given the strong negative and the scarce positive results for this problem without additional assumptions, we augment our algorithms with buffers that can store and rearrange a limited number of items. This setting interpolates naturally between the fully online case (no buffer) and the fully offline case (a buffer large enough to hold all items). We show that algorithms equipped with reasonably sized buffers can achieve strong guarantees for personalized $k$-value instances, i.e., instances in which each agent assigns at most $k$ distinct values to items. In particular, we construct allocations that are EF1 at every time step and EF at most time steps, using a buffer of size linear in $k$ and in the number of agents. Our approach relies on novel combinatorial arguments and on constructing a sequence of envy-free matchings that allocates most items. Finally, we extend our results to general additive valuation functions, with a dependence on the largest per-agent ratio between two values of the same sign, and we also identify limitations of our approach via impossibility results on the use of buffers with smaller size.
- [729] arXiv:2607.01539 (replaced) [pdf, html, other]
-
Title: MMAO-Cls: Metabolic Multi-Agent Optimization for Joint Feature Selection and Classifier TuningComments: 8 pagesSubjects: Neural and Evolutionary Computing (cs.NE); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
This paper studies whether the Metabolic Multi-Agent Optimizer (MMAO) can act as a credible outer-loop optimizer for classification model selection. We propose MMAO-Cls, a mixed-space realization in which each agent jointly encodes a binary feature mask and classifier hyperparameters, while private energy, communal budget, role drift, and lifecycle turnover are mapped to the accuracy-complexity tradeoff of wrapper learning. The implementation is strengthened by deriving feature-budget adaptation from feature-information priors and by regularizing validation reward with both subset compactness and train-validation overfitting gap. We evaluate MMAO-Cls on seven standard tabular benchmarks with three seeds each and compare it against RandomSearch, GA-lite, PSO-lite, and an endogenous no-sharing ablation. On the aggregate validation objective, MMAO-Cls ranks second ($0.9433$) behind GA-lite ($0.9446$). On held-out test performance, it reaches mean score $0.8882$, improving over RandomSearch ($0.8808$) and GA-lite ($0.8857$), remaining close to PSO-lite ($0.8874$) and the no-sharing ablation ($0.8900$), while using the most compact mean held-out feature subset among all compared methods (feature ratio $0.4881$). Pairwise tests show that these margins are not yet statistically significant. The results support MMAO's applicability to classification and compact mixed-space search, while the contribution of communal sharing to held-out performance remains unresolved.
- [730] arXiv:2607.01551 (replaced) [pdf, html, other]
-
Title: MMAO: Mechanism and Stability Analysis of Metabolic Closed-Loop MetaheuristicsComments: 12 pagesSubjects: Neural and Evolutionary Computing (cs.NE); Multiagent Systems (cs.MA)
This paper studies the Metabolic Multi-Agent Optimizer (MMAO) at the framework level rather than at the implementation or benchmark level. The central question is whether the metabolic resource loop of private energy, communal budget, role drift, and lifecycle turnover has a framework-level interpretation beyond narrative metaphor. We introduce a generic MMAO state model that abstracts away domain-specific move operators while retaining the resource bookkeeping that defines the framework. Under mild bounded-gain and bounded-spending assumptions, we establish boundedness and nonnegativity properties for private energy, communal budget, role state, and active population size. We then characterize three endogenous behavioral regimes of the loop: contraction under sustained resource deficit, reinvestment under surplus communal accumulation, and search redistribution under heterogeneous marginal returns across agents or subgroups. The analysis is intentionally conservative. It does not claim global convergence of the full adaptive system, universal superiority over specialist optimizers, or a complete stationary characterization of the resulting process. Instead, it identifies which internal regulation properties are generic consequences of the loop and which remain implementation specific. A compact mechanism-validation package on representative continuous and discrete MMAO realizations provides supporting empirical evidence for this reading, but is not intended to replace a full benchmark study. The analysis provides a bounded, regenerative, resource-regulated interpretation of MMAO, while leaving implementation-level performance questions open.
- [731] arXiv:2607.02797 (replaced) [pdf, html, other]
-
Title: RCOA Extension and ApplicationsComments: 11 pages, 9 figures, multimediaSubjects: Systems and Control (eess.SY)
The Relaxed Convex Obstacle Avoidance (RCOA) formulation is the first approach to enable a fully convex optimal control problem (OCP) for obstacle avoidance. Convergence analysis of RCOA yields an analytical framework that defines a unique characteristic: the ability to maintain obstacle avoidance (OA) efficacy even when obstacles reside beyond the controller's prediction horizon. In this paper, RCOA is extended to three-dimensional environments and apply it to Unmanned Aerial Vehicle (UAV) navigation. Furthermore, the formulation is enhanced to incorporate vehicle geometries, moving beyond point-mass representations to enable collision avoidance between 3D objects. Numerical simulations demonstrate that RCOA provides computational performance on par or exceeding state-of-the-art methods. Notably, RCOA is demonstrated to enable a Nonlinear Model Predictive Controller (NMPC) to execute aggressive maneuvers through narrow passages with reduced prediction horizons, ensuring real-time feasibility at frequencies exceeding 30~Hz.
- [732] arXiv:2607.03836 (replaced) [pdf, html, other]
-
Title: When Simpler Is Better: Evaluating Translation Pipelines for Medieval Latin ManuscriptsComments: 17 pages, EACL Findings 2027Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Despite remarkable progress in machine translation, Vision Language Models (VLMs) struggle on historical manuscripts, a domain that stresses core Natural Language Processing (NLP) capabilities: low-resource transliteration, archaic vocabulary, and noisy input signals. We present a systematic framework for evaluating the full image-to-translation pipeline on medieval Latin manuscripts, a setting in which scribal shorthand, ligatures, and parchment degradation expose failure modes that are invisible in clean-text benchmarks. Benchmarking on the CATMuS Latin dataset reveals a specialization gap: domain-specific Optical Character Recognition (OCR) models reduce character error rate by up to 4.3$\times$ compared to general-purpose VLMs, despite operating at orders of magnitude fewer parameters. We introduce the Interpres-Parallel-Corpus (IPC), a novel dataset comprising 1,383 aligned manuscript image lines, transcriptions, and expert translations, to our knowledge the first of its kind for medieval Latin. Our experiments uncover a complexity paradox: the simplest pipeline, a specialized OCR model feeding directly into a VLM, outperforms all multi-component variants. Adding retrieval-augmented generation (RAG) or post-OCR correction introduces prompt saturation and error propagation that degrade aggregate translation quality. These findings offer both a new benchmark and practical guidance for deploying translation systems in low-resource historical settings.
- [733] arXiv:2607.04688 (replaced) [pdf, html, other]
-
Title: URSA: Chemistry-Aware Benchmark for Utilitarian Retrosynthesis AssessmentBogdan Zagribelnyy, Ivan Ilin, Nikita Bondarev, Anton Morgunov, Arkadii Lin, Maksim Kuznetsov, Rim Shayakhmetov, Vladimir Aladinskiy, Alex Aliper, Alex ZhavoronkovSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Computation and Language (cs.CL)
Synthesis planning aiming to find pathways of reactions for a target molecule is one of the most important and challenging tasks in drug discovery. Recent progress has produced both specialized deep-learning retrosynthesis systems and general-purpose large language models, but objective comparison remains difficult due to the lack of flexible, chemically interpretable benchmarking protocols. In the current study, we are introducing the URSA (Utilitarian RetroSynthesis Assessment) evaluation framework that provides the opportunity to benchmark the synthetic routes not only from a formal perspective, such as convergence to commercially available starting materials, but also from a chemical plausibility perspective, mimicking the way expert chemists evaluate the reactions and routes. The study covers a comprehensive evaluation of both conventional end-to-end retrosynthesis solutions and LLMs for the synthesis planning task on a set of novel, diverse target molecules with undisclosed synthetic routes, which represent realistic tasks in the daily drug design routine. Although LLMs can support high-level strategic planning and even achieve SOTA performance in the OOD set, the Pareto frontier for estimated run cost and chemically plausible route solvability consists entirely of specialized retrosynthesis planners.
- [734] arXiv:2607.04926 (replaced) [pdf, html, other]
-
Title: Input Pathways Shape Few-Shot, Not Zero-Shot, Binding in Tiny Transformers: A Fully-Enumerable StudyComments: v2 withdraws the first few-shot factor: the compared pathways are one function class, so sharing was never manipulated, and the residual effect is the injected state's initialization scale, which vanishes under prefix injection (new appendix). Also withdraws the above-chance transient and the double dissociation (a mean of per-seed maxima). Readability and endpoint invariance stand. 18 pagesSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
How does the way information reaches a transformer -- as symbolic tokens, a clean per-factor "oracle" code, or an entangled perceptual vector -- shape whether it binds that information compositionally? We study ~6-10K-parameter transformers on finite factored worlds enumerated exhaustively, so every measurement covers the whole input space and the informative routes are information-matched (exact Bayes ceiling 1.0). Four findings. (1) Endpoint invariance: on held-out binding queries no informative route reaches converged zero-shot composition -- each ends at or below chance, so the failure reflects inductive bias under a lookup-sufficient objective, not missing information. It also holds with the state in its own sequence position rather than added to every token. (2) Readability governs few-shot binding: when a small fraction of the held-out type is leaked into training, sample efficiency tracks code readability. The least readable entangled code is worst (-0.147 pooled, under both schemes); a dimension-matched control and a graded readability sweep isolate readability from input dimension. (3) The injection site, not the pathway: v1 read the remaining ordering among readable routes as pathway parameter sharing; that is withdrawn. The compared pathways are block-decomposable, hence one function class, so sharing was never manipulated; the ordering comes from the initialization scale of the broadcast-added state and vanishes when it gets its own position (+0.057 -> -0.006, 20 paired seeds/cell). (4) Failure anatomy: symbolic routes lose the answer at the readout; index routes mis-bind -- the answer stays decodable, yet an input intervention shows the output tracks the wrong slot. The central positive claim is readability; the others bound what a pathway comparison can establish. Code, manifests, and per-seed logs are released.
- [735] arXiv:2607.05019 (replaced) [pdf, html, other]
-
Title: Fusion-Free Multimodal Classification with Heterogeneous Deep EnsemblesSubjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
In multimodal classification, late-fusion approaches classify concatenated modality-specific features extracted by unimodal neural networks. When modality imbalance is pronounced, various regularization techniques have been proposed to balance the learning process and overcome the inferior performance of late-fusion networks. In contrast, this work demonstrates that multimodal data can be effectively classified without any explicit modality fusion, using deep ensembles of unimodal networks. We systematically compare deep ensembles to late-fusion networks at equal parameter count and show that ensembles consistently outperform state-of-the-art late-fusion methods designed to address modality imbalance when trained from random initialization. This advantage also holds over intermediate-fusion techniques we evaluated and over hybrid methods that combine unimodal and multimodal classifiers. We propose and empirically validate a method for selecting the number of models per modality in an ensemble, avoiding computationally expensive exhaustive search. Under extreme modality imbalance and small ensemble sizes, the heuristic indicates that ensembles of unimodal models trained solely on the stronger modality are preferable. As the ensemble scales up, incorporating models from the weaker modality becomes beneficial. Both predictions align with our empirical findings. However, the heuristic becomes less reliable when the diversity of ensemble members is low. To systematically explore the challenges of optimizing multimodal models, we propose a synthetic multimodal framework that allows control over both the number of modalities and their predictive strength. Our findings are consistent across synthetic and real-world datasets. Finally, by fitting scaling laws to bimodal datasets, we estimate the asymptotic performance of ensembles.
- [736] arXiv:2607.07738 (replaced) [pdf, other]
-
Title: REFORGE: A Method for Benchmarking LLMs' Reverse Engineering Capabilities in Decompiled Binary Function NamingComments: 10 pages, 5 figures. Published in the proceedings of the 23rd International Conference on Applied Computing 2026 (AC 2026), Lisbon, 24-26 October 2026. Best Paper AwardJournal-ref: Proceedings of the 23rd International Conference on Applied Computing 2026 (AC 2026), Lisbon, Portugal, 24-26 October 2026, IADIS Press, pp. 202-211, ISBN 978-989-8704-82-5Subjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Programming Languages (cs.PL)
Large language models (LLMs) are increasingly applied to reverse-engineering tasks, and recent threat-intelligence reporting shows them operating inside live offensive-security workflows. Claims about their capability, however, outpace our ability to measure it. Existing benchmarks for LLM-assisted binary analysis treat the construction of function-level ground truth as a solved pre-processing step and report accuracy without disclosing how many functions were reliably evaluable. We argue that the principal obstacle to fair evaluation is not model capability but the reliability of binary-to-source alignment under compiler optimization. This paper presents Reforge, a provenance-tracked pipeline that constructs function-level ground truth from C source through compilation, DWARF and syntactic extraction, alignment, and decompilation, and that operationalizes alignment uncertainty as an eight-gate confidence funnel with three-tier stratification. On a controlled micro-benchmark, high-confidence yield falls from 87.2% to 65.9% across optimization levels, and unpaired comparisons overstate optimization-induced performance decay through survivorship bias. A proof-of-concept evaluation of seven contemporary LLMs on function naming demonstrates the validity of the concept and generally motivates an uncertainty-aware benchmarking practice.
- [737] arXiv:2607.07824 (replaced) [pdf, html, other]
-
Title: CPM-MultiAgent: A CPM-Grounded Appraisal Framework for Simulating Dynamic Patient Emotional Evolution in Healthcare DialogueJingyao Cai, Shuaijun Liu, Xiaowan Li, Yutong Guo, Abdul Rehman, Qintian, Thomas Dolby, Sue M Green, Xiaosong YangSubjects: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI)
Large Language Models (LLMs) enable patient simulation for healthcare communication training, where learners practice responding to clinical information, concerns, and emotional reactions. However, existing patient simulation and affective dialogue systems often represent patient emotions as static attributes or focus on recognizing emotions and generating empathetic responses. They provide limited support for modeling how a simulated patient's emotional state evolves in response to a healthcare professional's communication. Changes in patient emotion constitute interactional feedback that learners must recognize and respond to across dialogue turns. We draw on the Component Process Model (CPM), a psychological theory that explains emotion as a dynamic process shaped by event appraisal. We propose CPM-MultiAgent, a CPM-grounded multi-agent framework for dynamic patient emotion simulation in healthcare communication training. The framework uses an emotion taxonomy informed by medical literature and expert input, and updates emotional states through healthcare-related trigger identification, structured CPM appraisal, and emotion-state integration, accounting for the patient's profile, clinical context, and prior emotional state. We evaluate the framework through student-led conversations and controlled trigger cases, combining nursing-expert assessments, LLM-judge comparisons, component ablations, backbone checks, and student feedback.
- [738] arXiv:2607.12753 (replaced) [pdf, html, other]
-
Title: RFMSR: Residual Flow Matching for Image Super-ResolutionSubjects: Computer Vision and Pattern Recognition (cs.CV)
Generative image super-resolution (ISR) methods typically sample from a standard Gaussian prior and transport the sample to the data distribution, where the low-quality (LQ) input serves only as an external condition. However, the LQ image already carries rich structural information about the target high-quality (HQ) image, which makes a distribution approximating the LQ image a more suitable starting point, an idea that has been validated by a series of diffusion-based restoration methods. In contrast, recent flow-matching-based ISR methods still adopt the standard Gaussian prior and thereby discard such structural priors; moreover, existing single-step acceleration techniques, such as distillation and consistency training, either depend on an additional teacher or sacrifice the model's multi-step inference capability. In this paper, we propose Residual Flow Matching for Image Super-Resolution (RFMSR), which defines the source distribution as a Gaussian centered at the LQ latent, so that the flow starts from the LQ image rather than from pure noise and preserves the structural priors already present in the input. To achieve high-quality single-step generation, we further propose a simple yet effective training strategy that combines the supervision loss on the single-step output with the conditional flow matching (CFM) loss, enabling high-quality restoration with an arbitrary number of steps without resorting to distillation or other acceleration techniques. Extensive experiments show that RFMSR achieves competitive results against state-of-the-art methods. The source code is available at this https URL.
- [739] arXiv:2607.20289 (replaced) [pdf, html, other]
-
Title: Courteous Anticipation: Improving Long-Lived Task Planning in Persistent Shared EnvironmentsComments: 9 PagesSubjects: Robotics (cs.RO); Artificial Intelligence (cs.AI)
We consider a task planning scenario in which robots sharing a persistent environment are assigned tasks one at a time from a held-out sequence. Standard task planners, lacking foresight of future tasks and inconsiderate of others' constraints, solve each task in isolation, leaving terminal states that increase future cost for all; side effects that compound over lengthy task sequences. To reduce cost over the sequence, a robot must anticipate how its actions now may impact performance on future tasks for all robots sharing the environment. Therefore, we present courteous anticipatory planning, wherein a model-based planner proposes candidate plans and selects the one that jointly minimizes immediate cost and aggregated expected future cost across all robots, estimated via independent per-robot learned estimators. This factored formulation avoids combinatorial joint rollouts and supports modular deployment: adding a robot requires only training its own estimator. We evaluate in two persistent PDDL domains, a home environment with robots that have similar capabilities but different responsibilities, and a restaurant environment where robots' distinct capabilities create states that other robots lack the capability to resolve. During lengthy task sequences, our planner reduces total cost by 10.4% versus myopic and 4.5% versus selfish anticipatory planning in a two-robot home environment and by 20.5% and 6.7%, respectively, in a three-robot restaurant.
- [740] arXiv:2607.21118 (replaced) [pdf, html, other]
-
Title: The Second LoViF 2026 Challenge on Real-World All-in-One Image Restoration: Methods and ResultsXiang Chen, Hao Li, Jiangxin Dong, Jinshan Pan, Xin Li, Hongbo Ding, Junpeng Jiang, Xingyu Qiu, Yilian Zhong, Yuxiang Chen, Shibo Yin, Zixuan Huang, Yushun Fang, Xilei Zhu, Yahui Wang, Chen Lu, Xiaodong Zhou, Qingyue Cao, Changwei Gong, Jingyun Liu, Xingchen Yi, Hansen Shi, Ruiyi Liu, Jirui Xie, Tao Liu, Wenzhuo Ma, Hongzhen Li, Yongyong Chen, Zheng Zhou, Jingyong Su, Jie Liu, Haijin Zeng, Cheng Li, Peishuai Zha, Ziyi Wang, Jian Tang, Yan Chen, Long Bao, Heng Sun, Jiyuan Zhang, Shuai Liu, Wei Ding, Chengjun Guo, Yibin Huang, Xiaotao Wang, Dongqing Zou, Lei Lei, Xiaofeng Wang, Xiao Liu, Yulin Wu, Yuhan Zhao, Shurui Peng, Chao Ren, Yu-Kai Wang, Kosuke Shigematsu, Asuka Shin, Rong-Lin Jian, Cheng-Jun Kang, Jin-Hui Jiang, Jialin Zhou, Kuo Yuan, Songyu Zhang, E B Benson, Ashfaq Hussain, Pruthvikanth AC, Qirui Chen, Jinyuan Chen, Jun Zhang, Xu Zhang, Xuhui Cao, Jiaqi Ma, Laibin Chang, Yuchun Miao, Yichu Xu, Yuanzhi Yao, Shi Chen, Yuning Cui, Huan Zhang, Lefei Zhang, Saeed Ahmad, Ik Hyun Lee, Jun Young Park, Ji Hwan Yoon, Shangquan Sun, Behrooz Nobahar-Moghanlou, Majid Edalatjou, Karim Shahi-Niyar, Ruibo Zhang, Dexiang Hong, Xinyan Liu, Shengeng Tang, Weidong ChenComments: ECCV 2026 Workshops; this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV)
This paper presents a review of the second LoViF Challenge on Real-World All-in-One Image Restoration. The challenge aims to advance unified image restoration under diverse real-world degradation conditions, including blur, low-light, haze, rain, and snow. It provides a common benchmark for evaluating the restoration accuracy, robustness, and generalization capability of models across multiple degradation categories within a unified framework. The competition attracted 158 registered participants, and 20 teams were included in the final ranking after their submitted results were successfully reproduced and verified. This report provides a comprehensive analysis of the submitted solutions and corresponding results, highlighting recent advances in real-world all-in-one image restoration. The summarized methods and empirical findings reveal effective design strategies and establish an updated benchmark for future research in real-world low-level vision.
- [741] arXiv:2607.22117 (replaced) [pdf, other]
-
Title: Projection Pursuit CPCANet for Domain GeneralizationComments: This paper has been consolidated into the updated version of our work, available at arXiv:2605.05136Subjects: Computer Vision and Pattern Recognition (cs.CV)
Domain Generalization (DG) aims to learn representations robust to distribution shifts. Recent geometric alignment methods, such as CPCANet, extract domain-invariant structures through batch-wise Common Principal Component Analysis (CPCA). However, CPCANet suffers from rank-deficient covariance estimation due to the small-sample-size issue in mini-batch training. To address this limitation, we propose Projection Pursuit CPCANet (PP-CPCANet), a covariance-free framework that learns a global orthogonal basis on the Stiefel manifold and jointly optimizes it with network parameters via the Cayley transform. We further introduce a symmetry-breaking detached-median PP dispersion objective to extract common principal components (CPCs) with dense and robust optimization signals. Experiments on four DG benchmarks show that PP-CPCANet achieves SOTA performance while maintaining stable training.
- [742] arXiv:2607.22132 (replaced) [pdf, html, other]
-
Title: Majority Illusions in Large Random NetworksSubjects: Social and Information Networks (cs.SI)
An agent in a social network is under majority illusion when the majority among its neighbors differs from the majority in the whole network. We study analytically the probability of the occurrence of majority illusions on large random networks. We derive 0-1 laws for graph sequences that locally converge, using different types of random opinion-assignment functions. Our results show that in large random networks with uniform random opinion-assignment the probability of majority illusions goes to 0. However, when opinions depend on nodes' degrees the probability may go to 1. From a methodological point of view, our work illustrates how local convergence techniques from probability theory can be effectively applied to the understanding of opinion-related phenomena---such as majority illusions---in large networks.
- [743] arXiv:2607.23863 (replaced) [pdf, html, other]
-
Title: Polynomial-time computation of $\ell_p$-contraction fixed points for even $p$Subjects: Data Structures and Algorithms (cs.DS)
We give a $\text{poly}(d, p, \log(1/\epsilon))$-time algorithm that computes an $\epsilon$-approximate fixed point of any $\ell_p$-nonexpansive map $f : \mathcal{X} \to \mathcal{X}$, where $\mathcal{X} \subset \mathbb R^d$ is a convex compact set and $p$ is an even integer. This is the first algorithm with $\text{poly}(d, \log(1/\epsilon))$ runtime for any fixed $p \ne 2$. Our techniques are based on a computationally efficient version of Sion's theorem for non-compact minmax problems, and extend to more general total search problems that admit low-degree polynomial potentials.
- [744] arXiv:2607.24558 (replaced) [pdf, other]
-
Title: Proceedings of The First Reflection in Creative Experience (RiCE) WorkshopCorey Ford, Olga Sutskova, Samuel Rhys Cox, Sarah Sterman, Max Kreminski, Rosa Van Koningsbruggen, Anqi Wang, Ege Otenen, Karly Ross, Giulia Di Fede, Yinmiao Li, Salvatore Andolina, Marit Bentvelzen, Pan Hui, Nick Bryan-KinnsSubjects: Human-Computer Interaction (cs.HC)
Reflection and metacognition are central to the creative user experience. However, most HCI research on reflection focuses on clear, task-oriented goals such as to reflect on personal data or pedagogical outcomes. This contrasts with the open-ended and challenging to articulate goals of creative user experiences. For the first time, this workshop brings together interdisciplinary researchers, designers, educators, and artists across HCI, Cognitive Science, Design, AI, Learning Sciences, and Digital Art to examine reflection in creative interaction. The workshop will discuss themes, drawn from earlier discussions with HCI researchers and artists, on: how best to capture reflection in creative contexts, how to leverage the arts to support reflection for ethical change, and how to design creative AI that enhances - not hinders - critical thinking. By bringing interdisciplinary perspectives on reflection into discussion, the workshop will develop a guiding taxonomy for reflection in creative interaction to inform future creative practice and tool development.
- [745] arXiv:2607.29617 (replaced) [pdf, other]
-
Title: When Does On-Policy Interaction Help? Representational Tradeoffs in Value-Based Imitation LearningSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
Imitation learning (IL)---training an agent to replicate expert behavior from demonstrations---underpins applications from robotics to language model training. Standard approaches such as Behavior Cloning (BC) are known to suffer from compounding errors and performance plateaus, particularly when the learner cannot perfectly represent the expert's policy (as is typical, e.g., in distillation). Two interventions are widely understood empirically to improve performance: querying the expert interactively along the learner's own trajectories, and using value function estimation en route to generating a policy.
We investigate the nature of these improvements and their potentially surprising interplay. Our main finding is that expert interaction relaxes the representational demands on the learner: one only needs a model capable of realizing the expert's value function, bypassing the (often stricter) requirement of realizing the expert's policy itself. Concretely, we introduce OVI, an interactive on-policy IL algorithm that is statistically and computationally efficient whenever the learner can represent the expert's value function. We complement this with a negative result showing that interaction is necessary. Namely, without stronger assumptions beyond expert-value realizability alone, any offline IL algorithm must scale with the complexity of the expert policy class. Our findings bear out empirically. OVI outperforms offline policy-based (BC), interactive policy-based (DAgger), and offline value-based IL methods, with the largest gains when the learner network is substantially less expressive than the expert's. - [746] arXiv:2608.01570 (replaced) [pdf, html, other]
-
Title: Characterizing Treatment-Context Medication Evidence Across Clinic Notes and Structured EHR Medication HistoryComments: 8 pages, 3 figuresSubjects: Computation and Language (cs.CL)
Clinic notes may describe medication changes that are not reflected in structured electronic health record (EHR) medication history at the same encounter. Differences in terminology and documentation timing further complicate the interpretation of these discrepancies. We evaluated medication normalization in clinic notes and examined how semantic and temporal matching affected overlap with structured history. A note-grounded reference was constructed from large language model (LLM)-proposed mentions, deterministic canonical projection, and existing targeted correction records; random-audit records were summarized separately. On 5,360 reference mention rows from held-out patients, conditional canonical agreement increased from 81.25% with surface matching to 93.58% after lexical cleanup and curated aliases. Candidate-reference mention alignment was evaluated separately, yielding an F1 score of 0.6547. Across 27,552 reference rows, exact same-visit overlap was 16.71%. Overlap reached 90.22% when ingredient and broad-category matches were permitted at the same visit or within 30 days before or after the note. Incomplete mapping of note-derived labels contributed to residual nonoverlap in ontology-backed comparisons. Exact-label comparison captured only part of the related medication information across sources; broader matches do not establish equivalent medication use or treatment events.
- [747] arXiv:2608.02596 (replaced) [pdf, html, other]
-
Title: Staying Productive Under the Palm Trees: On Graded Coeffect Typing in the Tropical SemiringSubjects: Logic in Computer Science (cs.LO); Programming Languages (cs.PL)
We give a temporal reading of graded coeffects: tropical grades carry guards on infinite computation. To illustrate it, we first define $G_T$, a polymorphic calculus with guarded recursive types that guarantees productive infinite computation, in the form of hereditary head normalisation. The calculus supports stream computing via the Scott encoding, and embeds Nakano's later modality. We then introduce $I_T$, a graded intersection type system with guarded infinitary types for the pure {\lambda}-calculus, whose typability allows for an exact characterisation of hereditary head normalisation. We prove that the characterising typability predicate is $\Pi_0^2$, matching the recursion-theoretic complexity of the characterised property.
- [748] arXiv:2608.02612 (replaced) [pdf, html, other]
-
Title: BBOWP-Bench: Evaluating LLMs on Black-Box Optimization Word ProblemsComments: Accepted to NeurIPS 2026 (Evaluations & Datasets Track)Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
Formulating an optimization problem strongly affects final solution quality, yet good formulations usually require substantial expertise. Recent studies have explored automatically deriving optimization problems from natural-language descriptions, but existing benchmarks focus on settings where objectives and constraints can be written explicitly as mathematical expressions. Many practically important problems are naturally treated as black-box optimization (BBO) problems, in which only objective values are observable, and the functional form is unavailable. In BBO, search space design (as part of problem formulation) and optimization algorithm selection are crucial for problem-solving. Automating these processes with large language models (LLMs) remains challenging. We introduce Black-Box Optimization Word Problems (BBOWP), a novel problem setting in which a system must infer both a search space and an optimization algorithm from a natural-language description of a black-box optimization task. To support research on this setting, we establish the BBOWP Benchmark Suite (BBOWP-Bench), a dataset and evaluation framework for BBOWP. Each instance combines a natural-language problem description, an executable evaluation environment, and a human-designed reference formulation, allowing evaluation of both search-space design and algorithm selection. Using this benchmark, we evaluate LLMs on the joint formulation of search spaces and optimizers from natural-language descriptions across multiple BBO domains. We show that current LLMs vary their optimizer choice with the evaluation budget, though they sometimes struggle with search space design, particularly in identifying important variables and balancing their ranges when the problem description is less informative or the search space is highly problem-specific. Our code and dataset are available at this https URL.
- [749] arXiv:2608.07244 (replaced) [pdf, html, other]
-
Title: A constrained Onsager variational framework for parametric approximations of Willmore and Helfrich flowsSubjects: Numerical Analysis (math.NA)
We develop a constrained Onsager variational framework for parametric finite element approximations of Willmore and Helfrich flows. The weak curvature relation is used as a PDE constraint to express the first variation of the bending energy in curvature-vector form. Onsager's principle then combines this variation with dissipation based on the normal velocity, while an independent relaxed minimal-deformation-rate (relaxed-MDR) constraint selects the tangential velocity. The resulting mixed formulation admits a semidiscretization by continuous piecewise linear elements that satisfies an exact energy-dissipation law and exactly preserves area and volume when the corresponding constraints are imposed. The framework accommodates constant spontaneous curvature and applies to both closed surfaces and open surfaces with fixed boundaries under Navier or clamped conditions. We also propose a linearly implicit fully discrete scheme, prove its unique solvability, and present numerical experiments demonstrating convergence, energy decay, constraint preservation, and effective mesh redistribution.
- [750] arXiv:2608.09654 (replaced) [pdf, html, other]
-
Title: Hallucination-Free GUI Grounding via Regression-Free Layout-Aware MatchingSubjects: Artificial Intelligence (cs.AI)
GUI agents are shifting from metadata-dependent large language models to purely visual multimodal large language models (MLLMs) that operate directly on screenshots. The core task, GUI grounding, requires translating abstract user instructions into precise element coordinates. This task faces a persistent dual obstacle: conventional grounding models lack the semantic richness to interpret abstract instructions, while end-to-end MLLMs suffer from coordinate hallucinations caused by deficient fine-grained perception. We propose a regression-free framework where a frozen MLLM performs instruction parsing and a dedicated grounding model handles precise localization without learning any coordinate regression. A frozen MLLM first elaborates the abstract instruction into a structured visual description rich in layout cues. These descriptions are then fed to a novel Layout-Aware GUI Grounding Model, which performs regression-free localization by matching against layout-prior candidates, inherently suppressing hallucinations and avoiding expensive fine-tuning. The grounding model is trained with only Text/Icon binary labels, requiring no coordinate regression parameters. On ScreenSpot-Pro, our method achieves over 20% improvement in grounding accuracy over end-to-end systems; on Mind2Web, it raises success rate and element selection rate by more than 15%. These results demonstrate that decoupling instruction understanding from layout-aware localization effectively resolves the core challenges of GUI interaction.
- [751] arXiv:2608.15138 (replaced) [pdf, html, other]
-
Title: ReForge: Keeping ABR Algorithms Never Finished with Verified Large Language Model RevisionsSubjects: Artificial Intelligence (cs.AI)
Designing an ABR algorithm for a network scenario takes an engineer months, and large language models (LLMs) now do this work in hours. But either way, the algorithm is fixed once it is deployed. It fits the networks seen during its design and falls behind on the networks that arrive afterwards. We ask whether an ABR algorithm can keep being designed after deployment, revised as each new network arrives, without becoming worse on any network it already serves. We propose ReForge, which keeps a pool of existing ABR policies unchanged and lets an LLM revise only the rules that decide which policy chooses the next bitrate. When a new network arrives, the LLM reads how the current rules and every policy perform on the networks served so far and rewrites the rules. The rewrite is deployed only if a replay in simulation shows that it improves the served networks as a whole and makes none of them worse. A fast GPU simulator that replays every episode exactly makes this acceptance test cheap enough to run for every revision. We evaluate ReForge on nine real-world network datasets, with new networks arriving one at a time. The revisions raise the mean QoE from 1.43 to 1.79, past the best single policy at 1.66 and close to the 1.86 of selecting the best policy for each dataset, while the QoE on the networks served first does not drop. The gain also holds in a real video player. All code, data, and records will be available.
- [752] arXiv:2608.18389 (replaced) [pdf, html, other]
-
Title: Programmable Distributions of Semantics-Preserving Variants for Evaluating Code AgentsHasan Najib Mahmud, Shreya Gupta, Isha Chaudhary, Nathaniel Enis, Ravi Mangal, Gagandeep Singh, Corina PasareanuSubjects: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
Code agents are increasingly deployed to resolve real software issues, yet their reliability under superficial code variations remains poorly understood. We evaluate whether coding agents that repair repository-level issues remain reliable when the surrounding codebase is rewritten into a semantically equivalent form. We sample semantically equivalent variants of a repository from an explicitly defined distribution and compare the agent's resolve rate on them with that on the original. Different distributions reveal different aspects of an agent, such as its sensitivity to routine refactoring or to misleading code near the location of a fix. We develop VDL, a probabilistic programming language for expressing such distributions, whose programs randomly apply semantics-preserving transformations (SPTs) to a repository. For our evaluation, we use it to define a generic distribution that is not targeted at any particular failure mode. We evaluate two agentic scaffolds (mini-SWE agent and OpenCode), each backed by one of four frontier models (Claude Opus 4.5, Kimi K2.5, MiniMax M2.5, and Qwen 3.6-27B), on instances from SWE-bench Verified and SWE-bench Pro. Resolve rates drop by up to 6.7 percentage points, significantly in 6 of 16 configurations. No ranking of models by robustness holds across scaffolds, revealing a jagged robustness frontier.
- [753] arXiv:2608.18982 (replaced) [pdf, html, other]
-
Title: Monroe: A Molecular Foundation Model for In-Context Probabilistic InferenceComments: 40th Conference on Neural Information Processing Systems (NeurIPS 2026)Subjects: Machine Learning (cs.LG); Biomolecules (q-bio.BM); Quantitative Methods (q-bio.QM)
Bioassay activity prediction is often data-limited because drug-discovery datasets rely on time-consuming and expensive wet-lab experiments for data generation and evaluation. This challenge has inspired recent research into molecular foundation models (MFMs), which aim to encode general-purpose chemical knowledge into molecular representations that generalize well in data-constrained scenarios. This paper presents Monroe, a new MFM with several innovations over the existing state of the art: increased scale allowing pre-training on over 81 million molecules from the PM6 quantum chemistry dataset; improved graph representation of stereochemistry; improved training losses including conformer denoising and embedding decorrelation; improved multi-task learning; and the use of a prior-data-fitted model (TabPFNv3) for downstream in-context prediction. Our evaluations use a principled pairwise comparison framework that measures statistically significant performance differences. Across established Polaris benchmarks, Monroe matches or exceeds existing MFMs, while on activity cliff benchmarks, designed to assess utility for molecular discovery, it achieves significant improvements over prior methods. Finally, ablation and transfer experiments show that PFN-based downstream predictors also substantially improve two leading existing models, MiniMol and CheMeleon, yielding new state-of-the-art variants we call MiniMol_PFN and CheMeleon_PFN, suggesting that our downstream adaptation strategy generalizes beyond Monroe. Source code is at this http URL
- [754] arXiv:2608.19372 (replaced) [pdf, html, other]
-
Title: The Missing Touch: Sensing and Displaying Localized Contact Geometry Brings Teleoperation Closer to Human DexteritySubjects: Robotics (cs.RO)
A fundamental challenge in robotic teleoperation is enabling an operator to control a remote robot as effortlessly and intuitively as their own hands. Despite the growing use of teleoperation to collect demonstrations for training autonomous robot policies, teleoperated performance still falls significantly short of human dexterity, even for basic tasks. Here, we present evidence that a key factor contributing to this performance gap is the absence of localized tactile feedback. We equip a two-degree-of-freedom (DoF) bilateral force-feedback telemanipulator with a vision-based tactile sensor and a 32-pixel tactile display, allowing us to capture spatial contact patterns on the robot finger and render them on the operator's fingertip. In teleoperated button-discrimination and peg-rolling tasks, we demonstrate that reproducing spatially distributed contact information brings teleoperated movements closer to natural human behavior through fewer unnecessary motions and task completion steps, significantly reducing task completion time and producing more consistent demonstrations across trials and operators. We further show that these improvements in human demonstrations can propagate to autonomous policies: policies trained on tactile-enhanced demonstrations exhibited similar reductions in both task completion time and deviation from natural human behavior. Together, these results establish spatially distributed tactile feedback as an important component of dexterous robotic teleoperation and demonstrate its potential to improve safety-critical teleoperation and enable more humanlike demonstrations for training autonomous robot policies.
- [755] arXiv:2608.20344 (replaced) [pdf, html, other]
-
Title: From Survey Histories to Interpretable Digital Twins: Auto-Discovering Task-Specific Personas and Behavioral MechanismsComments: 13 pages. Accepted at SocialAgent: Second Workshop on Large Language Models for Social Reasoning and Simulation, NeurIPS 2026. Camera-ready version. Substantially revised title, abstract, methodology, and experimentsSubjects: Computation and Language (cs.CL); Computers and Society (cs.CY)
LLM-based "digital twins" aim to simulate how an individual would behave in new environments or respond to novel questions, given some representation of that individual's prior responses. We ask whether long survey histories can be transformed into compact, task-specific personas that remain predictive for unseen respondents and products within the same task while making simulated decisions more interpretable. We introduce an iterative auto-discovery pipeline that constructs persona representations and learns how persona evidence influences simulated responses. We evaluate the procedure on the 40-item Product Preferences--Pricing task of Twin-2K-500 (Toubia et al., 2025), learning from 20 respondents and 20 products and evaluating the frozen representation and mechanism on 30 unseen respondents and 20 unseen products. The holdout answers come from two survey waves, and every prediction is scored against each wave separately. For gpt-5.4-nano, replacing the full survey transcript with the learned channel representation and mechanism improves exact accuracy from 0.573 to 0.623 and from 0.558 to 0.626 in the two waves. Larger and cross-provider simulators (gpt-5.4, gpt-5-mini, qwen3-8b) using the full transcript do not achieve statistically significant improvements over the baseline in both waves, and none significantly outperforms the learned representation in either wave. This pattern shows that the learned representation produces consistent gains against human responses collected in both waves while retaining an explicit mapping from respondent evidence to predicted decisions.
- [756] arXiv:2608.21317 (replaced) [pdf, html, other]
-
Title: From Regulation to Implementation: A Critical Evaluation of LLM-Assisted Regulatory Compliance in IndustryComments: 10 pages, 4 figuresSubjects: Artificial Intelligence (cs.AI)
The European Union (EU) has emerged as a leading regulatory body in the development of sustainability and privacy regulations. While new regulation requirements vary, many include a documentation artifact to ensure compliance. Notably, the Ecodesign for Sustainable Products Regulation (ESPR) introduces Digital Product Passports (DPPs) for life cycle transparency, while the General Data Protection Regulation (GDPR) mandates Data Protection Impact Assessments (DPIAs) to mitigate privacy risks. Creating these compliance artifacts, however, is challenging. Industrial data, which often exists in heterogeneous formats and is scattered across company and supplier systems, is required for DPPs and can be difficult to extract into compliant DPP formatting. Furthermore, DPIA documents require interdisciplinary expertise and follow no standardized format, making development difficult for novel systems. To address the particular complexity of compliance artifact creation for both regulations, researchers have proposed the use of LLMs in the generation process; however, the impact of the aforementioned problems on the output of these systems is largely unaddressed. This work investigates the existing research gap by exploring how data extraction instructions and regulatory vagueness impact the quality and consistency of LLM-produced compliance artifacts. The resulting artifacts are evaluated by benchmarking different models against manually created ground-truth schemas. The results reveal that less strict guidelines, such as DPIA formatting, require higher context prompts to maintain consistency and completeness. Stricter guidelines, such as formatting for Digital Battery Passports (DBP), result in consistent results regardless of prompt context, but may lead to more hallucinations in the output
- [757] arXiv:2608.22228 (replaced) [pdf, html, other]
-
Title: Prompt-Based Abstention Fails Under Misleading Context: A Controlled Study of Small Frozen RAG ModelsComments: 19 pages. Published in AACL-IJCNLP 2026 (Main Conference). Code and data: this https URLJournal-ref: Proceedings of IJCNLP-AACL 2026 (Volume 1: Long Papers), pages 3368-3386Subjects: Information Retrieval (cs.IR)
Missing and misleading evidence are not the same problem in retrieval-augmented generation (RAG), but prompt-based abstention treats them alike. Small frozen models abstain when context is absent, not when it is misleading. We introduce GRAB-RAG (Graded Abstention Benchmark for Retrieval-Augmented Generation), a paired benchmark that tests the same questions across four context conditions (supportive, degraded, missing, and misleading) in Natural Questions and HotpotQA. In the misleading condition, we edit a gold passage to support a wrong answer and place it among other retrieved passages. We test five abstention policies on three small frozen models (3.8B-8B) across two English QA benchmarks. Models abstain reliably when evidence is missing, but under explicit abstention prompting still answer 41.6% of misleading questions, with 63% of those answers echoing the planted wrong entity. Chain-of-thought provides little additional benefit. A generator-side conflict check cuts the rate to 13.3% but discards many correct answers, while a natural language inference (NLI) verifier recovers that coverage but fails when parametric memory and the misleading passage agree on the same wrong answer. Prompt-based abstention asks whether context is sufficient, not whether it is correct. The two-pass checks are diagnostic baselines, not deployable mitigations.
- [758] arXiv:2608.23014 (replaced) [pdf, html, other]
-
Title: AnaDiffusion: Anatomically Compositional Latent Diffusion for Controllable 3D Brain MRI GenerationHuiwen Han, Lulin Liu, Bangya Liu, Yuanhao Cai, Nuo Chen, Xiaoqing Wang, Ziqian Xie, Chenyu You, Shuiwang Ji, Degui Zhi, Zhiwen FanSubjects: Computer Vision and Pattern Recognition (cs.CV)
3D brain MRI generation has made significant advances in medical imaging, simulation, and controllable anatomical analysis. However, existing generative models typically synthesize 3D volumes monolithically, often overlooking regional anatomical structures and limiting local controllability. To address these limitations, we introduce AnaDiffusion, an anatomically compositional latent diffusion framework that factorizes the generation process into distinct, anatomically meaningful regions, followed by part-to-whole assembly and global refinement. Our approach first trains part diffusion models to capture local structural priors. We then inject an assembled anatomical composite of the parts into the whole-brain latent representation and continue denoising. This mechanism enables the model to resolve global context while preserving the injected anatomy. AnaDiffusion thus produces explicit part assets and a globally coherent volume, enabling controllable part editing without subject-specific dense segmentation maps at inference. On a subject-disjoint ADNI test split, AnaDiffusion attains the lowest mean Fréchet distance (FID) in every evaluated region. In localized editing, it achieves substantially higher target transfer (paired MS-SSIM) than a segmentation-conditioned baseline, with off-target preservation close to the baseline's.
- [759] arXiv:2608.23642 (replaced) [pdf, html, other]
-
Title: AI Agents Push Humans Out of the LoopComments: Accepted to NeurIPS 2026, Position Paper TrackSubjects: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
AI agents pose significant risks as they are granted increasing autonomy. A commonly proposed solution is human oversight and keeping a ''human in the loop'', but this is not a simple solution: Not only do current approaches to AI agent design impede effective human oversight, but the cognitive capacities required for it are also themselves degraded by extended use of AI systems. This position paper argues that current approaches to the development and deployment of AI agent systems do not support effective human oversight -- they contribute to its degradation. To address this, a top priority in the advancement of AI agents should be supporting the situated goals and cognitive requirements of effective human oversight, treating the human needs of overseers at the same level of importance as AI agent capability. To put this idea into practice, we connect work on automation and human-computer interaction to AI agent processes, outlining design-level affordances and organizational protocols that (1) support overseers in exercising critical judgement and (2) counteract the skill atrophy that arises from extended use of automation. We urge developers and deployers to adopt these or similar approaches. Without explicit support for the cognitive demands of effective human-agent interaction, AI agent systems will continue to passively incentivize the degradation of the very human skills they rely on.
- [760] arXiv:2608.23718 (replaced) [pdf, html, other]
-
Title: The least-squares RBF-PU method for linear elasticity in the diaphragm geometrySubjects: Numerical Analysis (math.NA)
The human diaphragm is vital for respiration and its simulation can aid in further understanding complications arising in individuals that have been subjected to mechanical ventilation. The diaphragm is very thin and hence its numerical simulation poses a challenge for most numerical methods. In this paper, we compute the deformation of real diaphragm geometries, which are modelled as linearly elastic, and are subject to realistic boundary conditions. The method used is the unfitted least-squares radial basis function partition of unity method. The patches used are cylindrical and refinement is performed along the thin direction of the diaphragm. Both the unfitted nature of the method as well as the refinement technique make this method particularly well suited to solve problems formulated in thin geometries. We derive a theoretical proof showing that the method converges with high order to the solution of a well-posed linear elasticity boundary value problem.
- [761] arXiv:2608.29583 (replaced) [pdf, html, other]
-
Title: Drive the Thoughts: Runtime Monitoring of VLA Reasoning-Trajectory ConsistencySubjects: Software Engineering (cs.SE)
While machine learning models help autonomous vehicles (AVs) navigate complex environments, their black-box nature complicates validation and verification (V&V). Vision-language-action (VLA) models offer a unique opportunity: alongside trajectories, they generate an explicit chain-of-thought (CoT) explaining their decisions. We posit that this CoT can serve as a candidate specification to cross-check trajectories and expose unsafe behavior. To assess this, we curate DriveAlignBench, a dataset of 150 CoT-trajectory pairs from NVIDIA's Alpamayo 1.5 VLA, manually annotated for CoT reliability, CoT-trajectory consistency, and trajectory safety. We find that 33.3% of CoTs are unreliable, and among reliable CoTs, 26% have inconsistent trajectories. To detect these cases, we propose integrating a CoT-trajectory consistency check into a runtime monitor. This check is nontrivial because CoTs express open-vocabulary, scene-relative maneuvers, while trajectories are low-level ego-motion sequences whose semantics depend on road geometry and motion context. To bridge this gap, we develop a family of automated consistency monitors. The strongest, the lane-aware monitor, adds lane-relative evidence to the input of the judge model scoring consistency. This raises F1 across all four evaluated judges from 0.56-0.62 with raw waypoints to 0.65-0.82, showing that evidence representation improves detection across judges. Moreover, for reliable CoTs, inconsistency is associated with unsafe behavior: 27% of human-labeled inconsistent trajectories are unsafe compared to 3% of consistent ones, and the lane-aware monitor flags most of these unsafe cases. We release DriveAlignBench, the monitor implementations, and per-run results at this https URL.
- [762] arXiv:2609.00518 (replaced) [pdf, html, other]
-
Title: Learning Task-Specific Antibody Representations via Function-Aware MaskingComments: Accepted to MLCB 2026 (Oral)Subjects: Machine Learning (cs.LG); Biomolecules (q-bio.BM)
Antibody-specific language models pretrained via masked language modeling (MLM) learn representations that are critical for downstream sequence design and property prediction tasks. Yet, the corruption process itself is rarely leveraged as a source of inductive bias during pretraining. While preferentially masking complementarity-determining regions (CDRs) improves binding-related predictions, antibodies possess diverse biological priors over a variety of functions. Herein, we introduce function-aware masking, a family of pretraining algorithms that align mask placement with specific functional priors (e.g., from IMGT annotations or structure predictions) to shape the learned representation space. We show that these specialist masking strategies significantly improve performance on their respective objectives, yielding up to a 14% gain on structure-related tasks and up to a 5.9x improvement on CDR-related tasks. To further improve performance across multiple functional axes, we develop hybrid masking strategies that integrate multiple priors, balancing reconstruction over binding, structural, and biophysical objectives. Our results demonstrate that informed mask placement provides a parameter-free mechanism for imposing functional inductive biases in antibody language model training.
- [763] arXiv:2609.04141 (replaced) [pdf, html, other]
-
Title: Efficient Test-Time Adaptation through Human-Agent InteractionZora Zhiruo Wang, Apurva Gandhi, Rulin Shao, Aspen Chen, Jonas Mueller, Zhiqi Liang, Jett Chen, Michael Ryan, Qianou Ma, Luxi He, Zhoujun Cheng, Andre He, Seungone Kim, Jiayi Geng, Mingqian Zheng, Weiwei Sun, Zheyuan Zhang, Xinran Zhao, Yike Wang, Abe Hou, Liwei Jiang, Pang Wei Koh, Diyi Yang, Graham Neubig, Daniel FriedSubjects: Artificial Intelligence (cs.AI)
AI agents are trained on population-scale data to encode broad capabilities spanning those of many practitioners. Yet the artifacts they produce rarely meet the personal bar professionals need to stake their reputation on. On realistic, open-ended tasks where success criteria are heterogeneous and insufficiently documented, individual expertise lives precisely in the elevation and departure from the average. In practice, iterative human-agent interaction surfaces criteria that users cannot fully specify up front, yet apply repeatedly across tasks. We argue this cross-session interaction data is a rich, underused signal for closing the gap to individual expertise. In this work, we propose test-time adaptation through human-agent interaction (TAHI), which integrates these signals into agent context and weights, and crystallizes each user's training and evaluation criteria via an evolving rubric module. We adapt agents to 30 individuals in two high-utility domains, writing and visual creation, on a total of 600 tasks. Our agents improve solo task success by 4.5-20.9% within only tens of tasks. Meanwhile, our evolving rubric module serves as a scalable annotation tool, creating evaluation rubrics that catch 16.0-22.3% more failures than those from LMs or humans alone. While agents are adapted towards individuals, we show these personalized agents also produce improvements in success of up to 8.8% that generalize across users.
- [764] arXiv:2609.04218 (replaced) [pdf, other]
-
Title: A Governance Methodology Layer for AI-Assisted Software Development: Defect Taxonomy, Controlled Ablation, and a Test of Process-Over-CapabilityComments: v4: reports the outcome of the pre-registered test of C4 (Sec. 6.8); corrects references. 25 pagesSubjects: Software Engineering (cs.SE)
Autonomous coding agents produce output that passes syntactic checks -- compilation, type safety, CI -- at high velocity. Yet syntactic correctness does not imply semantic correctness: design boundaries, security invariants, and maintainability contracts remain structurally invisible to automated pipelines. This paper makes four contributions. First, we present a defect-class taxonomy grounded in five AI agent permission and governance modules, separating defects structurally detectable by static analysis from those requiring semantic review. Second, we describe a runtime-decoupled governance gate -- a file-based protocol that reads generator output and emits a structured verdict without API coupling, hence portable across generators. Third, we formalize methodology-as-code: expressing a verification protocol as a version-controlled, executable artifact whose two layers separate portable methodology from host automation. Fourth, we report a controlled ablation experiment (E-ablation, N=5 artifacts, 8-item independent ground truth) comparing harness-structured review against a token-matched unstructured prompt. The structured condition records 62% lenient recall against 50%, with 25% strict against 0%. A severity-grade differential reported earlier does not survive blind re-grading and is withdrawn (Sec. 6.6). An independent-session re-test with blind scoring does not replicate that contrast: the conditions differ by one strict hit in 24 (6/24 against 5/24), the structured aggregate again 25% and the unstructured 0% not recurring (Sec. 6.7). Both conditions miss document-quality defects identified by a human QA reviewer, indicating complementarity between structured AI review and human process inspection. A later pre-registered test varying model tier (184 runs) did not support process design as the dominant factor: under strict scoring the tier effect was larger (Sec. 6.8).
- [765] arXiv:2609.06351 (replaced) [pdf, html, other]
-
Title: A Theory of Information Architecture for Networked Decisions: Freshness, Locality, and CoordinationComments: 32 pages, 6 figures. Manuscript source and complete reproduction materials are available at this https URLSubjects: Systems and Control (eess.SY); Networking and Internet Architecture (cs.NI); Optimization and Control (math.OC)
A broader view of a network can help agents coordinate, but collecting that view takes time. When is a fresh local observation more useful than an older view of the whole network, and when should an agent look only partway across it? We study agents that repeatedly minimize a shared quadratic cost in an environment that evolves independently of their actions. An information architecture specifies what each agent observes and how old those observations are. Each agent uses a synchronized neighborhood snapshot, so extending its scope also ages the nearby readings used in the decision. We assume that the optimal action with complete information is affine in the state, and that the conditional mean of the current state equals the old state scaled by a common exponential decay factor. Using the classical Marschak--Radner decomposition, we derive the delay threshold beyond which fresh local information outperforms a complete but older global view. This comparison allows non-Gaussian states and nonlinear decision rules. We then determine how far an agent should look by balancing the benefit of additional observations against their greater age. In a Gaussian spatial model on a line, we derive the optimal observation radius: looking beyond the local node pays only when information travels more than twice the spatial correlation length in one coherence time. The preferred information architecture also depends on the decision being made. On a fixed 64-node ring, increasing only the reach of the decision weights from 0.7 to 6 hops shifts the optimal observation radius from one to five hops.
- [766] arXiv:2609.06383 (replaced) [pdf, html, other]
-
Title: From Reading Code to Reading Spec: A Verified Layer for LLM-Driven Codebase MaintenanceComments: Accepted by EACL 2027Subjects: Software Engineering (cs.SE); Computation and Language (cs.CL)
The rapid growth of LLM-generated code increases software complexity and the maintenance burden on engineers. While LLMs offer a potential automated alternative, this structural complexity hinders their ability to manage codebases directly. We introduce the Provable Representation Of Original Functionality (PROOF), which manages codebases indirectly via structured specifications. To enable full-lifecycle codebase management strictly through these specifications, PROOF abstracts codebase topology into a hierarchical natural-language representation. To establish absolute trust, the system proves semantic equivalence by reconstructing source code exclusively from this specification. This verified foundation drives maintenance requests, executing code modifications while synchronously updating itself to prevent semantic drift. Experiments on real-world repositories confirm the effectiveness of these specifications.
- [767] arXiv:2609.07729 (replaced) [pdf, html, other]
-
Title: Attributing Cohen's d: Training Data Attribution for Disease-Related Effects in Normative Age BiomarkersSubjects: Machine Learning (cs.LG); Quantitative Methods (q-bio.QM)
Normative age models are trained to predict chronological age in a nominally healthy cohort. In patients, the gap between predicted and chronological age is read as disease risk. We attribute the disease-related effect size of the age gap directly to individual training samples, rather than using a prediction-level loss as the attribution target. For the Cohen's d effect size, the resulting closed-form influence functional, validated against leave-one-out retraining, ranks training samples by their effect on held-out case-control separation. Across four diseases and two biomarker modalities in UK Biobank, removing the 10% most influential training samples raises held-out disease-related effect size in every seed. It more than doubles the metabolomic-age effect for type 2 diabetes and raises the brain-age effect for multiple sclerosis by roughly a third. Random removal leaves effect size flat even at 50% removal, confirming the gain comes from which samples are removed, not how many. Flagged subjects, missed by diagnosis-based exclusion, show elevated cardiometabolic markers the model never sees, consistent with subclinical contamination of the reference cohort. For type 2 diabetes, where the method gains most, the marker recovered is HbA1c, the standard measure of blood sugar control. We release our influence-function package pyinfluence.
- [768] arXiv:2609.08070 (replaced) [pdf, html, other]
-
Title: Local Spike Timing Enables Gradient-based Learning in Neural MicrocircuitsXiangnan Zhang, Jingxin Liu, Ranqi Lu, Jingyu Liu, Qunxi Dong, Fuze Tian, Lixian Zhu, Bin Hu, Björn W. SchullerComments: Version for the initial submission to Nature Computational Science, 13th Nov. 2026Subjects: Neural and Evolutionary Computing (cs.NE); Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
The brain uses discrete spikes for dynamic computation, yet in the fields of computational neuroscience and brain-inspired computing, how neural microcircuits (NMCs) solve temporal credit assignment using local spike timing remains a fundamental open question. Dominant spiking neural network (SNN) approaches circumvent this by approximating backpropagation through surrogate gradients, decoupling learning from biological spike timing. In order to integrate seemingly incompatible spike timing and continuous gradients, we reformulate temporal credit assignment as a state separation problem: extracting task-required components induced by historical perturbations directly from the current neural state. This enables an online feedback learning framework for NMCs through a gradient tunneling (GT) algorithm and the lead-lag expansion technique that derives credit assignment from local synaptic spike timing, while remaining compatible with ANN-SNN hybrid architectures. Experimentally, GT-trained NMCs excel at long-timescale evidence integration and noise-robust memory retention, and perform comparably to leading SNN online learning methods on real-world benchmarks with far fewer parameters. The proposed framework addresses the two-decade-old NMC feedback learning problem and suggests a computational theory for the brain's temporal credit assignment mechanisms.
- [769] arXiv:2609.11528 (replaced) [pdf, html, other]
-
Title: Mean Square Error Analysis of Stochastic Runge-Kutta IntegratorsSubjects: Numerical Analysis (math.NA)
We analyze the mean square error of stochastic Runge--Kutta integrators for overdamped Langevin dynamics whose potential is strongly convex outside a bounded region. A decomposition splits the local error into a mean-zero term and a smaller remainder, and the discrete Poisson equation turns their moments into a bound on the error of a time average. We carry this out for two stochastic Runge--Kutta integrators proposed by Yang \& Wang (2026), both of strong order $\frac32$ and weak order $2$, which evaluate the gradient of the potential and no higher derivative. We show that their laws approach the law of the exact solution at second order in Wasserstein-1 distance, up to a logarithm, uniformly in the number of steps. For a test function with bounded derivatives up to third order, we prove that the mean square error over $N$ steps with step size $h$ is $\mathcal O ( \frac{1}{Nh} + h^4 )$, which is the optimal order in the discretization. Experiments measure the strong and weak orders and the sampling bias on a nonconvex potential in $\mathbb R^2$, and compare the integrators on a diffusion model of CIFAR-10.
- [770] arXiv:2609.11728 (replaced) [pdf, html, other]
-
Title: Reproducibility in the Age of Agentic AI: Context Engineering at the Timescale of a CodebaseComments: 10 pagesSubjects: Software Engineering (cs.SE); Computers and Society (cs.CY)
Reproducible research practices are context engineering for AI coding agents. I argue that agents lower the cost of maintaining tests, commit histories, repository structure, instructions, and decision records while making their benefits immediate. Researchers remain responsible for verifying these artifacts and the scientific judgments they encode.
- [771] arXiv:2609.15664 (replaced) [pdf, html, other]
-
Title: Cloud-Edge LLM Collaborative Learning via Accelerating Personalized Bilevel OptimizationComments: We have fixed all issues, and submitted to IEEE TPAMISubjects: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
Amid the rapid advancement of physical-world intelligence, cloud-edge collaborative large language models (LLMs) have emerged as a promising roadmap for practical LLM deployment. However, existing cloud-edge paradigms struggle to balance global consensus with local personalization, which fails to satisfy the need for a unified knowledge foundation on the cloud and domain-specific adaptation at the edge. To address this, we propose a personalized bilevel optimization framework that formalizes cloud-edge LLM collaboration as a dual structure: the upper level optimizes edge-side personalization, while the lower level governs cloud-side knowledge transfer, reaching cloud-edge evolving in coordination. We then propose CIDERS: while the cloud performs knowledge transfer to the learnable backbone, the key lies in embedding global trajectories into each local personalization step via consensus-variate correction to reconcile personalization with consensus. We provide a comprehensive theoretical analysis, including a geometric characterization of the local trajectory and a full convergence guarantee, revealing an explicit trade-off structure between personalization and global convergence. Extensive experiments on Qwen2.5-3B and Qwen2-1.5B show CIDERS high performance and can significantly improve the accuracy (as much as 200\%+). Control-state diagnostics further show strong preservation of the local orthogonal component, while a separate cross-domain evaluation finds only a small matched-domain response advantage.
- [772] arXiv:2609.18078 (replaced) [pdf, html, other]
-
Title: Beyond Embedding Transfer: Architectural Interfaces, Iso-Parameter Scaling, and Preconditioner-Governed Consolidation in GrokkingComments: 26 pages, 7 figures, 19 tablesSubjects: Machine Learning (cs.LG)
Grokking, or delayed generalization after memorization, incurs substantial training steps across algebraic tasks. Cross-operator transfer from a grokked donor accelerates target acquisition, yet post-grokking AdamW continuation triggers test-accuracy relapse as unannealed updates drift across residual interfaces and decaying second moments amplify late gradients. Across depths one to three on modular arithmetic ($\mathbb{Z}_{113}$), non-Abelian $S_5$ conjugation, and real-sensory MNIST modular arithmetic, we separate representational transfer from preconditioner-governed consolidation across thirteen experimental batteries. First, co-transferring internal blocks accelerates one-layer and two-layer modular takeoff beyond boundary interfaces, while graded prefix pinning bounds depth-three drift under unannealed continuation. Second, an iso-parameter $S_5$ scaling grid separates attention head dimension from MLP width: widening two-head attention to dimension $64$ reaches $98.54\%$ cold accuracy at $427\mathrm{k}$ parameters and $85.95\%$ boundary-only transfer, while final-slot representations refine conjugation directly across blocks without an intermediate binary-product bottleneck. Third, Plateau-Gated Leaky Consolidation pairs validation-envelope step damping with leaky second-moment decay, reaching $84.85\%$ zero-leakage validation-gated risk-constrained accuracy across five pre-saturation two-layer $S_5$ suites while cutting step budget relative to hard anchoring. Fourth, transferring a grokked visual trunk and blocks to image-disjoint MNIST targets reaches $92.27\%$ and $90.76\%$ accuracy, prunes odd centroid harmonics, and reaches $90.42\%$ risk-constrained accuracy across $60/60$ branches without hyperparameter retuning.
- [773] arXiv:2609.23036 (replaced) [pdf, html, other]
-
Title: ResGraph: Cross-Model Graph-based Index Migration via Residual Reachability in ANN SearchBaoyuan Gu, Xiaoyao Zhong, Jiabao Jin, Peng Cheng, Wangze Ni, Haoyang Li, Haotian Li, Jingkuan Song, Heng Tao ShenComments: ICDE 2027 under reviewSubjects: Databases (cs.DB)
Approximate nearest neighbor search (ANNS) underpins large-scale vector retrieval in search, recommendation, and retrieval-augmented generation. Graph-based indexes have demonstrated state-of-the-art search performance for ANNS: they connect each corpus vector to a small set of nearby or navigationally useful vertices and answer queries by traversing the resulting graph. Because these edges are selected using construction-time distances, the graph index is tied to the embedding model; re-encoding a corpus with a new model may change distances and neighborhoods of the vectors, and reconstructing the graph for the new embedding vectors incurs substantial construction cost and delays deployment. When the embedding model changes, we observe a phenomenon in the old graph index that we call residual reachability: although derived from different models, the vectors describe the same underlying objects and often retain part of their similarity structure, which the old graph reflects as connectivity, leaving many exact new-model neighbors reachable within a few hops. Motivated by this observation, we develop ResGraph, an index-migration framework that exploits the residual reachability of the old graph index to construct the new graph index faster. ResGraph provides two migration paths. Res-Local performs parallel shallow expansion over the inherited graph index and screens second-hop candidates with packed position sign codes before exact evaluation. Res-Search uses hop-bounded beam traversal to explore beyond shallow expansion. Across eight text and image migrations, ResGraph attains up to 17.43 times speedup on constructing the new graph index than the fastest degree-matched reconstruction method while keeping competitive recalls.
- [774] arXiv:2609.23579 (replaced) [pdf, html, other]
-
Title: Contract-Based Decomposition of Temporal Logic Specifications for Networked Systems under Arbitrary PartitionsSubjects: Systems and Control (eess.SY)
Computational complexity is an inherent limitation of formal synthesis for networked systems, and decomposing the global specification into local ones relaxes this limitation at the cost of conservatism. Since the granularity of the partition governs this trade-off, it is reasonable to treat the partition as a design variable, which calls for local specifications that remain correct for every partition. To this end, this paper gives each agent a local specification, written as an assume-guarantee contract that the agent can establish from local information. We first derive a necessary and sufficient condition for these contracts to decompose the global specification under a given partition. Building on this, we then present a condition under which the decomposition is correct for every partition, so that the partition becomes a free design variable. For linear dynamics and signal temporal logic formulas with affine predicates, we further synthesize a controller for each coalition by a tube-based approach. Finally, simulations on a network of input-coupled tanks show how the choice of partition trades computational cost against conservatism.
- [775] arXiv:2609.27593 (replaced) [pdf, html, other]
-
Title: Hidden not Deleted: How Networks Suppress Entangled FeaturesSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
When a neural network stops reporting a feature, has it forgotten the feature or merely hidden it? Shared representations complicate the answer: removing one feature's direction can also damage another feature. We study this problem in a small autoencoder that compresses and reconstructs inputs, allowing individual feature contributions to be traced directly. With the representation fixed, restoring a single bias partially recovers target detection by exposing a signal hidden behind the output's activation threshold. Tests on target-absent inputs distinguish this recovery from interventions that simply increase output. The same bias intervention becomes less effective when the representation is allowed to change. We find that training only the other outputs can weaken the target's contribution, even without a loss on its own output. A conditional theoretical analysis explains how interference with retained predictions creates this pressure. Yet more flexible predictors still detect residual target information. These findings show how recovery depends on feature geometry and on which parameters are trained, and why a silent output is insufficient evidence of forgetting.
- [776] arXiv:2609.28610 (replaced) [pdf, html, other]
-
Title: UltraBench 2: Towards Robust Evaluation of Vision Foundation Models on UltrasoundComments: Accepted to Proceedings of the 6th Machine Learning for Health Symposium (ML4H 2026), PMLR Volume 346Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Benchmarking is an increasingly critical part of research in machine learning and the domains where it is applied, including healthcare. Yet, despite the steady development of new ultrasound foundation models in recent years, the development of well-designed benchmarks to evaluate them has lagged behind. This deficiency has led to fragmented and inconsistent evaluations of competing models, making it difficult to measure progress. To address this issue, we introduce UltraBench 2, a comprehensive benchmark with wide anatomical and task coverage, and a focus on standardization, reproducibility, and ease-of-use. Using this benchmark, we compare existing vision foundation models for ultrasound image analysis. Our analyses demonstrate that ultrasound-specific pretraining still leads on classification, but that state-of-the-art general-purpose models have drawn level on segmentation.
- [777] arXiv:2609.30210 (replaced) [pdf, html, other]
-
Title: The Alignment Illusion in Multimodal Large Language ModelsComments: Accepted to NeurIPS 2026Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Layer-wise visual-text similarity in Multimodal Large Language Models (MLLMs) is widely interpreted as evidence that the language model progressively integrates visual content into a shared representation space. This reading rests on the assumption that scalar alignment scores reflect content-level cross-modal interaction. To test this assumption, we apply controlled interventions to the visual stream. Across 13 MLLMs from five families spanning 0.5B to 72B parameters, replacing projector-output visual tokens with Gaussian noise sharply reduces task accuracy, yet four standard scalar measures (CKA, SVCCA, MIR, and the leading principal-angle cosine) fail to consistently separate the corrupted stream from the original. We call this failure the alignment illusion and trace it to the shared language-model pathway: anisotropic MLP down-projections pull visual and text tokens toward common output directions, producing weight-induced alignment. We introduce the principal-angle gap (PA gap), defined as the difference between the top two principal-angle cosines, to distinguish leading-direction dominance from the broader overlap observed with structured visual content. Under graded visual corruption, the PA gap tracks task accuracy more consistently than the scalar scores we consider; under a structured but irrelevant image, it further exposes regimes in which internal geometry and task accuracy come apart. Internal visual-text alignment in MLLMs is therefore best read as a geometric diagnostic of the visual stream inside the language model rather than a direct proxy for content-level cross-modal interaction, and is most informative when calibrated by controlled task evidence.
- [778] arXiv:2609.35075 (replaced) [pdf, html, other]
-
Title: ReCo: When to Relocate Sensor Kits under Deployment Constraints -- A NILM Case StudyComments: 9 pages, 3 figures, 6 tables. Accepted at the CONSTRAINT Workshop at IEEE BigData 2026. Updated manuscript and author affiliationSubjects: Machine Learning (cs.LG)
Many sensing tasks obtain training labels only by deploying instruments in the field. With a limited number of sensor kits, a collection deadline, and measurement downtime at every move, the collector must repeatedly decide whether to stay at the current site or relocate. We study this decision in non-intrusive load monitoring (NILM), which estimates the power drawn by individual appliances from a home's main meter and is trained on data from homes temporarily fitted with appliance-level sub-meters. In NILM, appliance usage varies with the appliance, season and climate, and the value of new data depends on how diverse the combinations of target operation and background load are. To address this, we propose a constraint-based relocation framework and instantiate it for NILM as ReCo (Relocation by Coverage gain). ReCo counts new operating regimes in a joint target--background feature space, forecasts each home's future gain from the data collected so far, and each night weighs the gain of staying against the gain of moving elsewhere after the downtime. In replayed deployments on the Plegma dataset under two kit counts and two downtime costs, ReCo outperforms fixed-dwell and count-based schedules and a threshold rule using the same metric in every setting. Its advantage is not explained by collecting more days alone and is consistent with allocating the days to more valuable homes and periods.
- [779] arXiv:2609.35348 (replaced) [pdf, html, other]
-
Title: From Data to Program: Fast & Direct Generative Program Inference from Empirical DataComments: 63 pages, 20 figures,Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Estimating probability densities from a finite set of samples typically requires dataset-specific model fitting. We introduce PRODiGI, a pretrained data-to-program model that infers an explicit, executable generative program in a single forward pass. Pretrained on synthetic datasets paired with their ground-truth programs, PRODiGI accommodates diverse generative families and data dimensionalities through template prediction and non-autoregressive program parameter decoding. Its inferred programs support direct sampling, density and score evaluation, and inspection independently of the pretrained model. We further introduce program-space fine-tuning, which refines differentiable program parameters by maximizing the likelihood of empirical data while keeping model parameters intact. Experiments show that PRODiGI-FT achieves better density and score metrics than existing pretrained models, while offering multi-fold speedups over its closest competitors. Program-space fine-tuning further reduces generation MMD by 90%. By turning empirical data into explicit, reusable programs, PRODiGI introduces a new direction for fast, interpretable tabular generative modeling.
- [780] arXiv:2609.35708 (replaced) [pdf, html, other]
-
Title: Mind the RefGAP: Correcting Reference Attention in Diffusion-Based Visual EditingSubjects: Computer Vision and Pattern Recognition (cs.CV)
Reference-guided diffusion editors struggle to faithfully reproduce user-provided references. We observe that many diffusion editors allocate limited attention to the reference. For example, in LoomVideo, edit-region queries assign less than 1% of their attention mass to the reference. We introduce RefGAP, a training-free correction that determines logit-offset magnitudes online at each layer from the reference-attention mass measured during the forward pass. Positive offsets to reference logits strengthen reference usage by edit-region queries, while negative offsets for keep-region queries limit reference-induced changes outside the edit. Two global coefficients control the correction; they are selected once on validation data from four development diffusion editors and held fixed. Across seven diffusion-based image/video editors, RefGAP improves identity fidelity in head swapping and face swapping, where identity takes priority; on some editors this raises pose error, which delaying the correction reduces. RefGAP achieves a fidelity-preservation trade-off comparable to separately tuned constant edit-side biases, without per-approach strength sweeps. Additional experiments on virtual try-on and background replacement evaluate transfer beyond identity editing.
- [781] arXiv:2609.37670 (replaced) [pdf, html, other]
-
Title: MeanFlowAdvantage: Stable Reward Fine-Tuning for Few-Step Average-Velocity GeneratorsSubjects: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Computer Vision and Pattern Recognition (cs.CV)
MeanFlow enables efficient few-step generation by predicting interval-average velocities, but this representation creates a mismatch for reward fine-tuning: existing advantage-based objectives are typically defined on instantaneous velocities or equivalent $x_0$-space predictions, whereas inference directly uses the learned average-velocity map. We introduce MeanFlowAdvantage, a signed advantage-weighted least-squares objective for average-velocity generators. Our key construction uses a shared, detached MeanFlow derivative correction to express the reward objective in prediction space while making rollout and reference regularization exact penalties on the average-velocity network deployed at inference. The resulting formulation preserves MeanFlow's native few-step sampler and provides a direct mechanism for transferring reward improvements to the deployed flow map. On SD3.5-Medium, MeanFlowAdvantage improves all eight reported metrics over the matched four-step MeanFlowNFT baseline and, with only four NFEs, matches or exceeds the 40-step DiffusionNFT baseline on six of eight metrics. The same objective also transfers to DNA promoter design, where it supports both teacher-free on-policy RL for a generator defined on a manifold and teacher-guided reward-graded distillation, with the latter yielding the lowest one-step Sei profile MSE among the compared configurations.
- [782] arXiv:2609.38540 (replaced) [pdf, html, other]
-
Title: When a Flatness Proxy Is Not a Function: Robustness Certificates and Training InterventionsComments: 10 pages, 3 figures, 1 tableSubjects: Machine Learning (cs.LG)
A valid curvature upper bound need not justify either a robustness certificate or an intervention on an intrinsic predictor property. We demonstrate this distinction for a last-layer relative-flatness proxy used in both settings. First, empirical-risk stationarity does not eliminate pointwise first-order loss terms: at a finite global empirical-risk minimum, the retained certificate expression underestimates a loss increase by over $210\times$. We derive a globally valid, gauge-invariant feature-space repair. Second, common-row softmax shifts preserve predictions and the exact contraction while making the proxy unbounded. Even standard reference-class choices double it on average relative to the centered representation. For a single fixed-feature example with at least three classes, scalar retuning generically cannot align the induced probability updates. Row centering gives the orbit-minimized bound and restores value and full-model gradient invariance under this symmetry. Across 45 paired one-step tests on algorithmic and image models, amplified shifts separate raw-regularized predictors while quotient-regularized predictors remain aligned. Long-horizon CIFAR-10 experiments show substantial, reversible suppression of generalization, while evidence for selective delay after memorization is less consistent. Together, these results show that validity as a curvature upper bound does not by itself justify either inversion into a robustness certificate or differentiation into an intrinsic training intervention.
- [783] arXiv:2609.39626 (replaced) [pdf, other]
-
Title: Parameterization method of reservoir properties for ensemble-based data assimilation using intermediate latent space of StyleGANSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Ensemble smoothers are the most successful and efficient techniques currently available for history matching. However, because these methods rely on Gaussian assumptions, their performance is severely degraded when the prior geology is described in terms of complex facies distributions (non-Gaussian). In this way, for these methods, we need to apply efficient parameterization techniques. Currently, the most efficient methods for performing parameterization are deep learning models. However, given the variety of existing deep learning models, studies have not identified which is most suitable for use with ensemble-based methods, although some important models had already been evaluated. Based on a recent literature review, the most promising models selected were VAE-GAN, Latent Diffusion, and StyleGAN models. As a novel aspect of this work, data assimilation with the second generation of StyleGAN (StyleGAN2) model was performed using the latent z-space and intermediate w-space, separately. They were applied in two 2D case studies: one categorical (three facies) and the other continuous. The results demonstrated that all three models are highly efficient, with the StyleGAN2 model standing out for generating samples with geological realism and achieving excellent data matching in the cases studied. Our findings show that performing data assimilation with StyleGAN2 using the intermediate space (w-space) yielded better results than the traditional application in the latent space (z-space). This is due to the fact that ESMDA uses linear updates and the w-space is much more linear and disentangled than the highly entangled z-space, thereby ensuring that the updated vectors remain close to realistic geological patterns. These results were validated using main geostatistical and history matching metrics.
- [784] arXiv:2610.00073 (replaced) [pdf, html, other]
-
Title: A Comprehensive Evaluation Framework for Conversational Home Energy Management SystemsComments: 25 pages, 7 figures, 9 tables. Re-set in the HUBS Lab paper layout (HUBS Lab Paper No. 2026-006); content unchangedSubjects: Human-Computer Interaction (cs.HC)
The growing complexity in home energy management (HEM) demands advanced systems that guide occupants toward informed energy decisions reflecting their background, preferences, and context. Large language model (LLM)-integrated HEM systems (HEMS) have demonstrated promise, but previous studies relied on single-turn or single-task evaluations with response accuracy as the primary metric. Whether such systems deliver effective interactions across the extended multi-turn dialogues typical of real-world use remains an open question. This study introduces a comprehensive evaluation framework of LLM-integrated HEMS derived from the Goal-Question-Metric methodology, organized across five categories: task performance, factual accuracy, interaction quality, control capability, and system efficiency. A total of 23 metrics across multi-turn conversations are proposed and an LLM-as-judge pipeline is employed to enable scalable automated scoring. Its reliability is validated against three trained human coders: after iterative rubric calibration, twelve of the fifteen LLM-scored metrics reached strong agreement (ICC >= 0.73), three of them perfect, while the remaining three exhibited near-zero variance in human scores and are instead reported via mean absolute error (0.04-0.28). To demonstrate the framework's effectiveness, 970 dialogues -- 16 scenarios and five personas -- were generated and evaluated across four conversational HEMS configurations spanning a sophistication gradient, from a vanilla LLM with raw energy data to a multi-agent HEMS. The framework distinguished the four configurations across multiple evaluation dimensions, revealing their respective strengths and weaknesses. This study contributes to conversational HEMS by providing a reproducible, multi-dimensional evaluation methodology that comprehensively assesses sustained, context-aware system performance.
- [785] arXiv:2610.00890 (replaced) [pdf, html, other]
-
Title: Cross-Benchmark Transfer from RL on Agentic Coding TasksComments: 15 pages, 2 figures, 4 tables. Accepted at ARRSI 2026, a workshop at AACL-IJCNLP 2026; also accepted as a non-archival paper at R2A-FM 2026Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
Coding agents often fail in the last mile: they build most of a feature but drop a requirement, test only the cases their implementation already handles, break behavior that was supposed to stay intact, or validate against an unchecked assumption. We ask whether reinforcement learning (RL) on expert-built agentic coding tasks closes this gap, and whether what the agent learns transfers beyond the training distribution. We post-train Kimi K2.7 Code, a 1T-parameter (32B active) open-weight mixture-of-experts model, with RL alone on 1,700 tasks: 1,000 repository tasks graded by hidden fail-to-pass tests and by pass-to-pass tests of existing behavior, and 700 terminal tasks graded by expert-written hidden verifiers. The reward is the fraction of target checks passed and drops to zero if any pass-to-pass test fails. One epoch of GSPO on a rank-32 LoRA adapter improves pass@1 on each of the six external benchmarks we evaluated, across three agent harnesses: SWE-Bench Pro (60.1 to 64.8), DeepSWE (31.0 to 43.4), Terminal-Bench 2.1 (67.4 to 82.0), Terminal-Bench 3 (1.4 to 12.1), Terminal-Bench 4 (0.0 to 7.6), and SWE-Marathon (5.0 to 25.0). Pooled over the five independent task sets (Terminal-Bench 4 revises Terminal-Bench 3), the improvement is significant (p < 0.001), and it remains significant on the three sets released after the training data was collected (p = 0.004); the model also improves under both harnesses never used in training. Median trajectories on DeepSWE and Terminal-Bench 3 are 24-35% shorter in agent steps. The base model's failed DeepSWE runs are mostly near-misses, and on the tasks the trained model newly solves, paired trajectories show it avoiding each of the four failure modes above.
- [786] arXiv:2610.02288 (replaced) [pdf, html, other]
-
Title: AdaptViT: Runtime-Adaptive Vision Transformer Deployment on Custom RISC-VVishnu PS, Ajay Kumar M, Yike Li, Robert Bogdan Staszewski, Deepu John (University College Dublin, Ireland)Comments: Accepted for publication at the 2026 IEEE 33rd International Conference on Electronics, Circuits and Systems (ICECS)Subjects: Hardware Architecture (cs.AR)
Deploying Vision Transformers (ViTs) on low-power edge devices is challenging due to high computational demands. Conventional pruning frameworks require a separate compiled binary for each sparsity level, increasing storage overhead and limiting runtime adaptability. This paper presents an end-to-end deployment pipeline that transforms pretrained ViTs into a single runtime-configurable binary, enabling dynamic compute-budget switching on embedded CPUs. This is achieved by restructuring generated C kernels with modified loop bounds and binary-mask control logic, allowing execution to switch across discrete sparsity levels via compact external configuration files. Compared to multi-binary deployment, the proposed runtime-adaptive approach reduces on-device storage by up to 4.86x, requiring only 163 MB for ViT-Base instead of nearly 800 MB. To maximize pruning efficiency, we introduce a hardware-aligned block pruning strategy for Multi-Layer Perceptron (MLP) layers. In addition, a custom ISA extension is proposed to exploit input-reuse patterns in linear projection kernels. On a Synopsys TRV32P3FX RISC-V processor, the full system achieves up to 2.8x speedup at 65% MLP and 50% attention-head pruning for ViT-Base. The ISA extension alone provides a 1.56x speedup and 33% lower inference energy, with a 24.7% area overhead in a TSMC 28 nm implementation.
- [787] arXiv:2610.02771 (replaced) [pdf, html, other]
-
Title: Nearly Optimal Fixed-Confidence Best-Arm Identification with 1-Bit FeedbackComments: To appear in Advances in Neural Information Processing Systems 39 (NeurIPS 2026, Spotlight)Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
We study fixed-confidence best-arm identification under strict 1-bit feedback constraints. At each round, the learner selects an arm and a query set, and receives only a single bit indicating whether the sampled reward belongs to that set. We consider a distribution-free finite-variance setting with arm-wise localization, where direct empirical mean estimation is no longer available and clipping becomes unavoidable. We first formulate a time-uniform 1-bit mean-estimation primitive based on randomized threshold queries and a clipped tail-integral identity. We then embed this primitive into candidate-challenger best-arm identification algorithms. A fixed-clipping algorithm gives a simple anytime $(\epsilon,\delta)$-PAC guarantee, while a phased adaptive-clipping algorithm matches the clipping level to the current resolution and yields a gap-adaptive sample complexity. We also prove a $K$-arm worst-case information-theoretic lower bound showing that the logarithmic penalty caused by finite-variance 1-bit feedback is intrinsic. This bound matches the leading dependence of the phased algorithm up to logarithmic factors.
- [788] arXiv:2610.02805 (replaced) [pdf, html, other]
-
Title: From TS-SUF-2 to TS-SUF-4: Practical Security Enhancements for FROST2 Threshold SignaturesSubjects: Cryptography and Security (cs.CR)
Threshold signature schemes play a vital role in securing digital assets within blockchain and distributed systems. FROST2 stands out as a practical threshold Schnorr signature scheme, noted for its efficiency and compatibility with standard verification processes. However, under the one-more discrete logarithm assumption, with static corruption and centralized key generation settings, FROST2 has been shown by Bellare et al. (in CRYPTO 2022) to achieve only TS-SUF-2 security, which is a consequence of its vulnerability to TS-UF-3 attacks.
In this paper, we address this security limitation by presenting an enhanced variant of FROST2, namely, FROST2+ which achieves the TS-SUF-4 security level under the same computational assumptions as the original FROST2. FROST2+ strengthens FROST2 by integrating additional pre-processing token verifications that help mitigate TS-UF-3 and TS-UF-4 vulnerabilities while maintaining practical efficiency. We show that FROST2+ can achieve TS-SUF-4 security not only under the same conditions as the original FROST2 analysis, but also when initialized with a distributed key generation protocol such as PedPoP. Our benchmark using ZCash's FROST library shows that the performance of FROST2+ is comparable to FROST2 and about 64-79% faster than FROST when precomputation is enabled. - [789] arXiv:2610.03713 (replaced) [pdf, html, other]
-
Title: What Should World Models Forget? Stratified Retention for Continual AdaptationComments: Accepted to NeurIPS 2026 Workshop on Continual World Models. Project Website: this https URLSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV); Signal Processing (eess.SP)
Continual learning treats any drop in performance on previously seen data as forgetting, a convention inherited from settings where the prediction target is stationary. World models violate this assumption: their target is the environment, which changes, so knowledge that was accurate when learned can later become false, and revising it is correct behavior rather than a defect. Non-stationary ground truth is well studied in the concept drift literature and in the temporal factuality of language models, but it has not been formulated for world models, which are distinctive in that they also encode structural knowledge, such as physical laws and object permanence, that must never be degraded. We argue that continual world models need retention stratified by invariance timescale, the time over which knowledge stays true: invariants should be refined but never degraded, while instance-level facts should be revised once the environment changes. We show that standard forgetting metrics, used alone, count correct revision as forgetting and so rank a frozen model highest. We propose differential retention, which reports two quantities separately: invariant retention, checked after every adaptation round, and revision latency, the number of observations a model needs to pick up a change.
- [790] arXiv:2610.04277 (replaced) [pdf, html, other]
-
Title: SyclKittens: A Tile Programming Model for Programmers and Coding Agents on Intel GPUsComments: 51 pages, 13 figures. Code at this https URLSubjects: Distributed, Parallel, and Cluster Computing (cs.DC)
New AI accelerators arrive before the kernels that make them fast, because peak kernel performance requires architecture-specific expertise in operand pipelines and data-movement techniques. Coding agents can now write, compile, and tune kernels on their own, so they could greatly accelerate kernel development and optimization. What agents produce depends on the interfaces they are given. These interfaces may expose a machine's raw capabilities or encode the known-good methods for exploiting its hardware features efficiently. We test the performance impact of these two interfaces on Intel GPUs through controlled experiments with five coding models on three kernels. With raw SYCL and execution feedback, Opus 4.8, the strongest of the three models tested with Claude Code, writes a single-GPU GEMM kernel reaching only 57.5% of Intel's tuned oneDNN library. We present SyclKittens, a hardware-aware tile programming model for Intel GPUs that encodes these known-good methods, so its operations place operands in matrix-engine layouts, prefetch through the L1 cache, and communicate over the fabric that links GPU stacks into a node. With SyclKittens under the same agent, task, and feedback budget, the agent reaches 82.1%, showing that encoded methods turn hardware capabilities into performance. Workload-specific policies stay programmable, and engineers and agents jointly refine these schedules in SyclKittens to build a kernel suite that reaches approximately 96% of oneDNN in geometric mean across GEMM shapes. The suite runs Llama-3.1-8B inference 1.59 times faster than this http URL on one GPU and up to 2.91 times faster than a matched multi-GPU decode path on Intel's oneCCL. SyclKittens is open source and available at this https URL.
- [791] arXiv:2610.05503 (replaced) [pdf, other]
-
Title: Lightweight Semantic EEG Foundation Model for Frozen Cross-Disorder TransferComments: Withdrawn pending administrative review of publication-related requirementsSubjects: Machine Learning (cs.LG); Signal Processing (eess.SP); Neurons and Cognition (q-bio.NC)
Large-scale EEG foundation models have demonstrated promising transferability across neurological disorders, but often require millions of parameters and substantial computational resources. In this paper, we present the Universal Semantic EEG Foundation Model (USE-FM), a lightweight EEG foundation model that learns transferable neural representations through self-supervised signal reconstruction on the Temple University Hospital EEG Corpus (TUEG). After pretraining, the encoder is frozen and evaluated on two clinically distinct downstream tasks, abnormal EEG detection (TUAB) and epileptic seizure recognition (TUEP), using a unified frozen-transfer protocol against recent EEG foundation models, including LUNA-Base and CBraMod. With only 1.46 million parameters, approximately one-fifth the size of existing models, USE-FM achieves competitive overall performance, including strong sensitivity and F1-score on TUEP (SEN $75.00 \pm 14.14$, F1 $70.37 \pm 4.01$), while maintaining competitive performance on TUAB (AUC $85.24 \pm 5.61$). Beyond downstream classification, latent representation analysis using $k$-means clustering together with PCA and t-SNE demonstrates that USE-FM learns organized semantic EEG representations comparable to substantially larger foundation models. These results suggest that large-scale self-supervised pretraining enables lightweight architectures to learn transferable semantic EEG representations, providing a computationally efficient foundation for cross-disorder analysis and future clinical decision support in neurological disorders.
- [792] arXiv:2610.06549 (replaced) [pdf, html, other]
-
Title: Proof-Grounded Patient-Specific Clinical Explanations from Knowledge-Graph ReasoningSubjects: Artificial Intelligence (cs.AI)
Clinical decision support can lack auditable links from observations and encoded knowledge to conclusions and recommendations. Downstream of a frozen, training-free clinical knowledge-graph reasoner, the CKG Clinical Explanation Engine converts patient inference states and disease knowledge into typed facts, explicit rule traces, provenance-linked conclusions, and policy-licensed recommendations. Separate availability, representation completeness, and disease-specific activation distinguish observed zero-activation and partial evidence from unobserved evidence. Language-generation evaluation remains future work. Across six usable workbooks (11,289 patients, 33,867 patient-disease traces, 1,715,928 feature-evidence rows), IG-range validity, knowledge provenance, cross-sheet numerical fidelity (203,202/203,202), and exported fact/report field presence (33,867/33,867) reached 100.00%. Availability-representation consistency reached 99.29% (1,703,811/1,715,928), with 12,117 disagreements across five systematic feature-cohort patterns. Observed zero-activation and partial instances numbered 111,488 and 226,880, respectively. A separate seeded 30-patient end-to-end audit ran failure-free, passing all prespecified trace, licensing, provenance, and state-consistency checks. Paired comparisons (90 traces, fixed inherited diagnostic scores and policy catalogue) showed the availability-only baseline changed every report but no action sets. State collapse changed 20 action sets and removed 10 urgent-review occurrences. Rule replay reproduced all 40 action changes from reclassifying partial evidence as unobserved. The bounded validator detected all 450 single-fault mutants. Structural and implementation auditability were established. Clinical correctness and utility remain unestablished.
- [793] arXiv:2610.07018 (replaced) [pdf, other]
-
Title: When to Rethink: Learning Multi-Perspective Self-Verification for Vision-Language ModelsZiquan Zhu, Hanruo Zhu, Si-Yuan Lu, Morris Yu-Chao Huang, Yicheng Lin, Wei Han, Tianlong Chen, Mingyuan Wu, Hanchao Yu, Gaojie Jin, Lu Liu, Bo Sun, Tianjin HuangComments: The authors request withdrawal of this preprint. We have obtained additional experimental results and are substantially rewriting the manuscript to incorporate these findings and expand the analysis. No replacement version is currently availableSubjects: Artificial Intelligence (cs.AI)
Vision-language models (VLMs) have achieved strong performance in multimodal reasoning, yet they remain prone to generating plausible but incorrect answers. Self-verification offers a practical way to improve answer reliability without relying on external judges, but existing methods typically depend on a single verification criterion or fixed prompt, resulting in incomplete and unstable reliability estimates. We first systematically analyze how verifier capability and prompt design affect verification performance. Our findings show that stronger verifiers provide more reliable judgments, while verification performance is highly sensitive to prompt choice, with no single prompt consistently dominating across tasks. Guided by these findings, we propose \texttt{MOTIVE}, a \textbf{M}ulti-View Self-Verificati\textbf{O}n wi\textbf{T}h Rel\textbf{I}ability-Guided Selecti\textbf{VE} Rethinking framework for reliable multimodal reasoning. \texttt{MOTIVE} evaluates each candidate answer from complementary verification perspectives and learns a correctness-aligned reliability score through correctness-grounded multi-view verification learning. During inference, this score governs an accept-or-rethink decision, allowing reliable answers to be returned directly while uncertain ones trigger history-guided rethinking. Extensive experiments across diverse multimodal benchmarks and VLM backbones demonstrate that \texttt{MOTIVE} consistently outperforms strong self-verification and self-correction baselines. Further results show that reliable verification improves accept-or-rethink decisions and reduces unnecessary reasoning turns, enabling more reliable and efficient self-verification without an external judge.
- [794] arXiv:2610.08650 (replaced) [pdf, html, other]
-
Title: Fast Non-Parametric Heteroscedastic Imitation Learning With Geometric PriorsMaximilian Mühlbauer, Arne Sachtler, Markus Knauer, Cem Küçükgenç, Yanlong Huang, Alin Albu-Schäffer, João SilvérioSubjects: Robotics (cs.RO)
When learning probabilistic policies from human demonstrations, data-efficient learning and fast adaptations to new scenarios are key requirements. One popular way to achieve intuitive and reliable adaptations is through non-parametric, typically kernel-based, methods. However, existing solutions either fail to account for the geometry of manifolds common in robotics, limiting data efficiency, or, when geometry-aware, provide unreliable uncertainty estimates or require retraining to adapt. We propose a non-parametric approach leveraging geometric priors in scenarios of data scarcity and heteroscedastic uncertainties for probabilistic modeling. We utilize the method to formulate policies based on time or robot state, where non-separable diagonal kernels allow capturing uncertainty relations between degrees of freedom for same-sized in- and outputs. Fast updates, requiring less than 3 ms for a trajectory involving both position and orientation are possible through an optimized formulation. Our approach supports both manifold-valued input and manifold-valued output with large orientation changes. Using task parameterization, adaptation to different object poses is easily possible. We evaluate the approach on a set of toy examples and on real robot manipulation tasks both in autonomous execution and in shared control scenarios.
- [795] arXiv:2610.10850 (replaced) [pdf, other]
-
Title: Similar Predictive Fit but Different Latent Dynamics: Characterizing Learned Dynamical Structure in Personalized Models of Brain DisordersComments: Withdrawn pending administrative review of publication-related requirementsSubjects: Machine Learning (cs.LG); Signal Processing (eess.SP); Neurons and Cognition (q-bio.NC)
As AI models move toward clinical decision-making and personalized treatment, understanding \emph{what} a model learns is important beyond predictive accuracy alone. We investigate whether personalized latent dynamics reveal clinically associated differences even when predictive fit is similar. A lightweight CNN--Transformer EEG foundation model pretrained on the Temple University EEG Corpus (TUEG) extracts segment-level representations. Using the Temple University Epilepsy Corpus (TUEP), representations are mapped to a shared latent-state space, and sparse multinomial logistic transition distributions (mLTD) are fit independently to each subject to obtain personalized transition-dependency graphs $W_n$. Analyses include $n{=}198$ subjects (99 epilepsy / 99 non-epilepsy). At $k{=}4$, epilepsy subjects exhibit substantially denser learned dependency structure ($p{=}1.1\times10^{-7}$), with the same pattern at $k{=}6$ (19.90 vs. 13.46; $p{=}5.2\times10^{-5}$). Graph-derived features provide moderate group discrimination under 5-fold subject-wise cross-validation (AUROC 0.68 at $k{=}4$; 0.65 at $k{=}6$). In contrast, held-out log-likelihood is nearly identical between groups at $k{=}4$ ($-0.992$ vs. $-0.991$; $p{=}0.95$), with similarly matched next-state prediction (AUROC 0.855 vs. 0.861; $p{=}0.54$). Thus, similar predictive fit does not imply similar learned dynamics: groups can be comparably predictable while differing substantially in the internal dynamical structure learned by personalized models. This distinction motivates evaluating learned structure alongside predictive performance in personalized clinical models.
- [796] arXiv:2610.11028 (replaced) [pdf, html, other]
-
Title: Synchronization Reveals Market Structure That Classical Randomness Tests Cannot DetectComments: 33 pages, 7 figures, 6 tables. v3: adds out-of-sample HAR baselines and an economic-value analysis, uses closed-bar information timing, and corrects forecast statistics affected by opening-time alignment of multi-timeframe data. Submitted to The North American Journal of Economics and FinanceSubjects: Computational Engineering, Finance, and Science (cs.CE)
Whether asset prices contain structure beyond a random walk is classically assessed with variance-ratio tests of the random-walk hypothesis. Synchronization measures from physics, notably the Kuramoto order parameter applied across trading timescales, have been proposed as an alternative diagnostic of market state, but they have been validated on single asset classes and never compared with the variance-ratio test on the same data. We compute a cross-timeframe Kuramoto coherence measure and a heteroskedasticity-robust Lo-MacKinlay variance ratio on identical windows for 61 instruments spanning cryptocurrencies, foreign exchange, metals, equity indices and single-name equities. The two diagnostics are statistically independent. We explain this analytically: coherence decomposes exactly into a direction-agreement channel and a volatility-agreement channel, and under sign-symmetric returns the direction channel is independent of the variance ratio and of every other statistic built from squared returns; simulations confirm the prediction. Coherence tracks the directional cleanliness of recent price moves in all five asset classes, whereas the variance-ratio test cannot reject the random walk for most of the panel. Held to finance benchmarks with strict information timing, however, coherence adds almost nothing to heterogeneous autoregressive volatility forecasts out of sample, and trading rules built on it earn no gross return outside cryptocurrency, where trading the coherence direction at high coherence breaks even at about five basis points per round trip. Failing to reject weak-form efficiency therefore does not rule out directional structure, but the structure that synchronization reveals is mostly descriptive and economically thin.
- [797] arXiv:2610.12392 (replaced) [pdf, html, other]
-
Title: Tiling 3D Space by (Translates of) a Single Polycube is UndecidableComments: 10 pages, 2 figures, 2 tables, 15 snippets. Added rotation/reflection results and corresponding tableSubjects: Computational Geometry (cs.CG); Combinatorics (math.CO); Metric Geometry (math.MG)
We prove co-RE-completeness, and thus undecidability, of the following problem: given a single (connected) polycube, decide whether it tiles 3D Euclidean space by translations. The same result holds if we also allow rotations and/or reflections. We reduce from Wang tiling using the decorated two-prime Sudoku construction of Greenfeld and Tao and a cyclic encoding adapted from OpenAI's 3D aperiodic tile, and apply a reduction of Kim to make the prototile connected (via faces). Dimension three is optimal: translational monotiling is known to be decidable in $\mathbb{Z}^2$ and for a single (possibly disconnected) polyomino in $\mathbb{R}^2$.
- [798] arXiv:2610.12654 (replaced) [pdf, html, other]
-
Title: Symbolic-Permutation-Based Direct Sparse Jacobian Assembly for Accelerated Power-Flow AnalysisComments: 11 pages, 3 figuresSubjects: Systems and Control (eess.SY); Performance (cs.PF); Numerical Analysis (math.NA)
The repeated assembly of the sparse Jacobian matrix constitutes a primary computational bottleneck in the Newton-Raphson power-flow analysis. Through an algebraic operator equivariance analysis of the Jacobian matrix, this paper formalizes the reduction process and proposes a direct sparse Jacobian assembly method based on symbolic permutation. By pre-applying the permutation to the admittance matrix during a fast one-time symbolic phase via its sparse prefix-sum property, the runtime bus-type projection overhead and indirect mapping caches are entirely eliminated. Consequently, complex Jacobian differentiation and global matrix assembly fuse into a single operation, yielding a single-pass, branchless algorithm that streams values directly into contiguous memory. The proposed algorithm is implemented in the open-source RustPower solver. Benchmarks on IEEE test systems and PEGASE 9241 demonstrate that the proposed assembly is 28.1-37.9 times faster than the MATPOWER-style baseline and 8.9-9.7 times faster than the in-house Schäfer-Braun-inspired assembly (V1); on PEGASE 9241, it outperforms optimized traditional C++ assembly routines by 3.72 times in cold construction and 2.59 times across iterations. End-to-end, the iteration-loop solve time with KLU is 1.16-1.97 times faster than the native C++ LightSim2Grid and up to 95.5 times faster than pandapower, while leaving numerical results and convergence behavior unchanged. The method extends directly to general problems where a sparse Jacobian inherits its structure from a static network operator, including transient stability analysis, EMT simulation, and OPF Hessian assembly.
- [799] arXiv:2610.13672 (replaced) [pdf, html, other]
-
Title: Inject a Turn, Not a Token: Learning Where to Redirect LLM Reasoning for Self-CorrectionSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Large language models struggle to reliably identify and correct errors in their own reasoning. Unlike prior training-based approaches that fine-tune an LLM to generate revisions or feedback, we learn where to intervene in failed rollouts and determine how to redirect them. To choose where to intervene, we train a 130K-parameter MLP head on a frozen encoder using only final-answer correctness. For redirection, we find that under a fixed generation budget, the tested inline injections fail to reliably redirect reasoning, whereas presenting the same instruction in a new user turn changes the model's course. The resulting correction framework, UPTURN (User-Prompt TURN injection at a learned position), requires no LLM fine-tuning: it retains a selected prefix, injects an instruction as a new user turn, and regenerates the suffix. Across four reasoners and six math benchmarks, UPTURN improves average accuracy in every setting over a 32-rollout baseline. It outperforms self-correction and reflection-token injection baselines under matched correction budgets, transfers across reasoners and to GPQA-Diamond without retraining, and complements GRPO and self-correction methods.
- [800] arXiv:2610.15020 (replaced) [pdf, html, other]
-
Title: PATTERN Can Be Solved from Memory: Testing Graph Neural Networks on Unseen PatternsComments: 48 pages, 6 figuresSubjects: Machine Learning (cs.LG)
PATTERN asks graph neural networks to find a planted subgraph in a new host graph. The hosts change, but the benchmark draws just 100 patterns and reuses them across training and test splits. A lookup that extracts these patterns from the training labels achieves 96.6% class-weighted accuracy on the official test split, above every published result, without fitting a model or consulting the generator. On unseen patterns, the same lookup scores 50.6%. A matcher that also uses the generator reaches 99.2% on the official split. Among five graph networks we retrained, GRASS and PPGT go from the top two on the official split to the bottom two on unseen patterns. Each loses 8.1 percentage points even when familiar and unseen patterns are matched in size. Logistic regression on two-hop node statistics scores above every run of the five networks on unseen patterns. For CKGConv, the advantage follows the training patterns: it reappears with each of three independently drawn pattern libraries and disappears when every training and validation graph gets a new pattern. Generator-based inference also scores 82.3% on CLUSTER, compared with a published best of 81.6%. On Minesweeper's released board, four of eight planned chains produced samples within the time budget; none converged, but their pooled predictions reached 99.86% ROC AUC, above the published best of 99.55%. PATTERN's official split tests its intended task of finding familiar patterns in new graphs. We recommend reporting the same checkpoint's score on unseen patterns alongside it.
- [801] arXiv:2610.16473 (replaced) [pdf, other]
-
Title: Some complete $ω$-powers of a one-counter language in the difference hierarchyJacques Duparc (UNIL), Olivier Finkel (IMJ-PRG (UMR\_7586)), Dominique Lecomte (IMJ-PRG (UMR\_7586))Subjects: Logic in Computer Science (cs.LO); Computation and Language (cs.CL); Formal Languages and Automata Theory (cs.FL); Computer Science and Game Theory (cs.GT); General Topology (math.GN); Logic (math.LO)
Fix a finite alphabet $\Sigma$ having at least two letters, and natural numbers $n,k\geq 1$. We denote by $D_n(\mathbf{\Pi}^0_k)$ the class of differences of $n$ subsets of the Cantor space $\Sigma^\omega$ in $\mathbf{\Pi}^0_k$, and by $D^-_n(\mathbf{\Pi}^0_k)$ the class of complements of sets in $D_n(\mathbf{\Pi}^0_k)$. We prove, for any natural numbers $n,k\geq 1$, the existence of a finite alphabet $\Sigma$ having at least two letters and a finitary language $U_{n,k}$ over $\Sigma$ accepted by a one-counter automaton such that the $\omega$-power $$U_{n,k}^\infty:=\{ w_0w_1\cdots\in\Sigma^\omega\mid\forall i\in\omega ~~w_i\in U_{n,k}\}$$ is $D_{2n-1}(\mathbf{\Pi}^0_k)$-complete. We also prove a conditional result for the dual classes $D^-_{2n}(\mathbf{\Pi}^0_k)$: under the assumption that there exists a finitary language $V_{n,3}$ accepted by a one-counter automaton, such that its $\omega$-power $V_{n,3}^\infty$ is $D^-_{2n}(\mathbf{\Pi}^0_3)$-complete, there would also exist, for any integer $k > 3$, a finitary language $V_{n,k}$ accepted by a one-counter automaton, such that its $\omega$-power $V_{n,k}^\infty$ is $D^-_{2n}(\mathbf{\Pi}^0_k)$-complete. Finally we add another conditionnal result about Wadge degrees of rational $\mathbf{\Delta}^0_2$ omega powers, which completes the recent results of [FGL26].
- [802] arXiv:2610.19095 (replaced) [pdf, html, other]
-
Title: Similar Models, Broken Odds: Where Luce's Axiom Fails in Chatbot ArenaComments: 20 pages, 5 figures, 10 tables; v3: adds a test on human summarization preferences and a calculation of votes needed for a release decision; clearer abstract, introduction and transfer results; appendix regroupedSubjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Chatbot Arena gives each language model one Bradley-Terry score, which fixes how often any model should beat any other. Choice theory predicts where this fails: near-copies are compared on the few details in which they differ, so votes between them are more decisive than their scores imply, exactly for the comparison a lab cares about, a new version against the old one. On 1.8 million Arena votes, votes between versions of one model are 1.37 times as decisive as Bradley-Terry predicts, between sizes of one generation 1.26 times and between models of one lab 1.12 times; a release a year later points the same way, and the effect survives tie, prompt, topic and timing checks. How alike two models' answers are worded and formatted tracks the effect. Plain scores leave the odds between relatives about 9% too close to even on held-out votes; a correction with four to six parameters roughly halves this and predicts held-out votes better, mostly by modeling ties, while the leaderboard order barely moves; frozen and carried to new releases, its gains in head-to-head odds are too small for our test to establish. With language-model judges and three-answer menus, a near-copy takes votes mostly from its twin when choices are pooled over prompts (log-odds shift +0.10, 95% interval 0.05 to 0.16), less so within a prompt.
- [803] arXiv:2610.20579 (replaced) [pdf, html, other]
-
Title: PingpongDuo: Agile Humanoid Table Tennis Doubles with Safe Team CoordinationHaoran Jiang, Yichao Zhong, Kai Zhang, Cheng Liu, Yixuan Pan, Zhenjie Yang, Jingbo Wang, Tianyu Li, Li ChenComments: 9 pages, 6 figures. Project page: this https URL . Minor updateSubjects: Robotics (cs.RO)
Unlike singles, humanoid table tennis doubles requires teammates to strike in strict alternation while sharing a confined workspace. Rapid repositioning can compromise striking accuracy, while uncoordinated teammate movements introduce collision risks. We present PingpongDuo, a hierarchical framework for agile humanoid doubles with safe team coordination. At the team level, a planner jointly selects positioning targets and uses learned risk estimates to filter risky proposals, preferentially preserving the current striker's target. At the individual level, separate footwork and striking skills combine agile movement with precise racket control. The skills are guided by human motion references, including footwork collected from human doubles play, and trained from states produced by one another to support rapid transitions. A racket-arm residual further refines strikes after movement using return feedback from a simulator calibrated with real ball-flight and contact measurements. In simulation, PingpongDuo improves the ball-return success rate by 13.2% over the state-of-the-art baseline, while achieving approximately 2 cm positioning error during clearance. In real-world experiments, a robot completes a 2.15 m clearance-and-recovery cycle in 2.4 s with a peak speed of 2.23 m/s, while two robots sustain more than 10 consecutive strikes in strict alternation.
- [804] arXiv:2610.21025 (replaced) [pdf, html, other]
-
Title: Learning to Read Out: Unembedding Dynamics in Language Model PretrainingComments: NeurIPS 2026. 10 pages main text, 51 pages total. Code: this https URLSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
When does a language model acquire a capability? Developmental studies usually answer this by tracking when models start producing the correct token. But that prediction passes through the unembedding matrix, or readout, which is itself still being learned during pretraining. We show that a model's hidden states can encode task information before its own readout expresses it, so timelines read from outputs conflate what the model encodes with what its readout expresses. To track the readout directly, we introduce parameter-trajectory crosscoding, which fits one sparse dictionary to every token's readout row at every checkpoint. Unlike crosscoders on activations, it needs only checkpoint weights. We use it to locate a narrow early window, 0.4-1.1% of the way through Pythia's pretraining, in which the readout reorganizes and WordNet word categories become decodable from its rows. By then, linear probes on the hidden states already decode task information that the model's own outputs do not yet show. To test whether such gaps depend on how fast the readout learns, we pretrain small models from scratch that differ only in the readout's learning rate. We find that raising it makes the readout develop earlier and the gap close sooner, with little effect on final loss. Together, these results imply that early in training, where many developmental transitions are dated, a behavioral date can reflect when the readout learns to express task information and not when the model learns it.
- [805] arXiv:2610.22063 (replaced) [pdf, html, other]
-
Title: Finite Volume Convergence for a Heat Equation with a Measure-Valued SourceSubjects: Numerical Analysis (math.NA); Analysis of PDEs (math.AP)
Past work with Tom Seidman analyzed the spatial convergence order of the numerical finite element method applied to time-dependent diffusion-reaction equations with a measure-valued source term. For smooth source terms, the numerical method has second-order accuracy, but the non-smooth source terms introduce a limitation of the order that depends on the spatial dimension. This work in honor of Tom extends the analysis to the finite volume method in its clas- sical two-point flux form and shows that the regularity of the source term reduces the convergence order to 1 in two and 0.5 in three spatial dimensions which is reduced from the nominal second-order and depends on the space dimension. Computational tests with a point source modeled by a Dirac delta distribution exhibit the behavior predicted by the analytical results.
- [806] arXiv:2610.23170 (replaced) [pdf, html, other]
-
Title: Chinese Decision: A Family of Language Models for Single-Pass Probabilistic Decision-MakingQiang Yan, Jiafeng Guo (Institute of Computing Technology, Chinese Academy of Sciences)Comments: 47 pages, 15 figures, 25 tables. Technical report v6. New since the previous arXiv version: unified 0.8B/2B/4B models with a typed span-extraction readout in the same weight file, the CJ-IE extraction mixture, and retention-cost analyses. Code: this https URL ; models: this https URLSubjects: Computation and Language (cs.CL)
Many applications need either a probability distribution over a few decisions or a set of typed spans from a text, and both are usually obtained by generating text. We present Chinese Decision, a family of Chinese models that answers both with a single forward pass and no generated token: candidate-letter logits give probabilities for Boolean, categorical, and ordinal decisions, and a small span head reading an echoed copy of the input gives typed character spans. Both readouts share one backbone and one merged LoRA, so a unified model is one weight file. We release decision-only and unified models at 0.8B, 2B, and 4B (Qwen3.5), trained for one epoch on 13.05 million decisions and, for the unified models, on CJ-IE, a 1.2-million-row Chinese extraction mixture. On the full 307,900-decision CJ-Bench, the unified 4B model reaches 86.86% accuracy, above every reported baseline and the decision-only 4B model (85.97%), and extracts spans on CJ-IE with micro F1 of 68.9. Retention has a price: against a 0.8B model continued on CJ-IE alone, which reaches 73.9 F1 but forgets CJ-Bench (71.02%), keeping decision ability costs 5.0 to 10.7 F1 points, shrinking with size. Gains of the unified models over decision-only ones are small, and all comparisons rest on single training runs with paired clustered bootstrap intervals.
- [807] arXiv:2610.23730 (replaced) [pdf, html, other]
-
Title: Synthetic Image Oversampling for Rare-Class Representation Learning in Deep Facial PhenotypingComments: BIBM 2026 ML4BMI WorkshopSubjects: Computer Vision and Pattern Recognition (cs.CV)
Recent deep facial phenotyping methods can support the recognition of genetic syndromes, but patient images are often scarce and unevenly distributed across disorders. Synthetic facial images offer a way to supplement the training data without collecting additional patient photographs. Their usefulness, however, depends on whether they improve recognition of real patients: generated images may introduce artifacts or repeat features already represented in the training set. It is therefore necessary to examine both how much synthetic augmentation to use and which disorders benefit. We investigate disease-specific synthetic oversampling using images from the Phenotypic Disease Image Database to augment a subset of the GestaltMatcher Database. We compare increasing augmentation levels in an image-based model, GestaltMatcher-Arc, and a multimodal model, GestaltMML, while keeping the real training and patient-disjoint evaluation partitions fixed. Repeated training runs assess overall and disease-specific accuracy, and an exploratory embedding analysis examines changes in class compactness and separation. Augmentation improves mean recognition accuracy in both training setups, although gains remain modest in GestaltMatcher-Arc and continue across the tested levels in GestaltMML. These aggregate improvements are accompanied by uneven disease-level responses, including reduced accuracy for some disorders. In the embedding analysis, disorders with fewer real training images tend to gain more class separation, but geometric changes do not consistently track classification performance. Together, the findings support synthetic oversampling for limited-data facial phenotyping and emphasize the need to assess its amount and benefit for individual disorders within each training setup.
- [808] arXiv:2610.23826 (replaced) [pdf, html, other]
-
Title: Self-Building LabsJiaru Bai, Yuchi Zhao, Yang Cao, Han Hao, Elia Savino, Miroslav Bogdanovic, Geyuan Xie, Xiaoman Guo, Derk H.W. ten Klooster, Simone Pilon, Timothy Noël, Varinia Bernales, Alán Aspuru-GuzikSubjects: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Robotics (cs.RO)
Self-driving laboratories (SDLs) accelerate scientific discovery by automating experimental execution and optimization. AI agents can increasingly propose new experiments, but existing SDLs can only run the procedures they were built for. Adapting them to a new experiment still requires extensive manual engineering, which makes laboratory flexibility the bottleneck for scaling discovery. We introduce the self-building laboratory (SBL), which designs, configures, and operates its own laboratory setup given a high-level scientific goal. SBL starts from whatever already exists, from an empty room to a working SDL, and builds the missing capabilities: the lab layout, its digital twin, robot skills, and experimental workflows. We demonstrate the system working across different robots and setups, scaling from a single workstation to laboratories with several robots working together. We validate SBL's end-to-end capabilities through a real-world quantum-dot synthesis experiment. By shifting human researchers from manually configuring and programming device automation to specifying scientific objectives, SBL makes scientific automation more accessible, flexible, and scalable. Project website: this https URL
- [809] arXiv:2610.24161 (replaced) [pdf, html, other]
-
Title: TorchTPU - A Hybrid Eager/Compiled Architecture for Native PyTorch Execution on TPUClaudio Basile, Michael Voznesensky, Gunhyun Park, Ben Wilson, Parker Holloway, Zhanyong Wan, Vlad Belous, Manish Khanna, Paul Ganssle, Kevin Gleason, Andrew Dame, Brendan Folie, Bhavya Bahl, Haifeng Jin, Pedro Goncalves Mokarzel, Arian Arfaian, Luke Boyer, Chi Zhang, Yiling Yuan, Adam Mainz, Aditya Patel, Huangze Tang, Jay Gan, Lizzie Liang, Will Froom, Tushar Agarwal, Karlo Basioli, Theotime Combes, Emilio G. Cota, Minos Papadopoulos, Tudor Pipernea, Yarong Mu, Qi Zhou, Yaoshiang Ho, Aleksey Vlasenko, Alexander Brauckmann, Ziyin Huang, Jack Oh, Jerry Hsu, Tanvi Bajpai, Jialei Chen, Wenjun Gu, Chintan Shah, Chris Chan, Sayce Falk, Chris Jones, Rob Torres, Kat Ko, Lee Howes, Freya Azad, Andi Gavrilescu, Joe Pamer, Bill JiaComments: 23 pages, 5 figuresSubjects: Artificial Intelligence (cs.AI)
TorchTPU is an open-source collaboration between Google and Meta that introduces a native and principled PyTorch backend for Google TPUs. It enables seamless out-of-the-box model execution and TPU migration while delivering industry-competitive inference and training performance through eager execution optimizations, this http URL integration, and native hardware acceleration.
- [810] arXiv:2610.24349 (replaced) [pdf, html, other]
-
Title: FlashRT-PAI: Fast Kernels Are Not Fast Pipelines for Physical AIComments: source code this https URLSubjects: Artificial Intelligence (cs.AI)
Vision-language-action (VLA), video-action (VA), and world action models (WAMs) form the backbone of modern physical AI systems, yet their real-time single-batch inference remains a defining bottleneck. Latency is dominated not by individual kernel computation, but by system-level overhead: thousands of short kernel launches, intermediate data movement across operator boundaries, and redundant recomputation across consecutive decisions. General-purpose graph compilers cannot address these costs, as they optimize individual calls in isolation and lack visibility into cross-call state semantics and global pipeline structure. We present FLASHRT-PAI, a low-latency inference engine built around model-hardware pipeline co-design rather than isolated kernel optimization. FLASHRT-PAI combines three core mechanisms: deep kernel fusion with launch overlap eliminates inter-operator boundaries and launch gaps; an execution contract formalizes the pipeline's stage DAG, update rates, and named buffers to enable cross-decision state reuse and selective refresh; and execution-state capsules enable fast snapshot, restore, and fork of persistent pipeline state. The system also supports drop-in acceleration of unmodified PyTorch models via parity- and latency-gated kernel replacement. Evaluation covers six policies on RTX 5090 and four on Jetson AGX Thor. At matched 16-bit precision, FLASHRT-PAI is 1.3-3.7x faster than official compiled paths. Combining pipeline co-design with reduced precision delivers 3.38-12.66x speedup over official eager implementations. In LIBERO-10, four-bit pi0.5 achieves 94.04% success on Thor and 93.32% on RTX 5090, compared with 92.4% for the measured official FP16 reference. FLASHRT-PAI is open source, distributed via Hugging Face Kernel Hub, and its opt-in pi0.5 kernels have been adopted into the LeRobot framework.
- [811] arXiv:2610.24451 (replaced) [pdf, other]
-
Title: Assessing the CAD Capabilities of General-Purpose AI Models: A Case Study of GPT-6 AstraComments: The paper contains 180 page long full report of CAD Capabilities of General-Purpose AI Models, and its 12 pages of summary. The project page is at: this https URLSubjects: Graphics (cs.GR)
Recent advances in general-purpose AI models and agentic systems have enabled them to perform some CAD tasks. A systematic assessment is now needed to establish what these systems can accomplish and how well their outputs meet engineering requirements. This paper presents an empirical study of GPT-6 Astra, a SOTA general-purpose AI model, across four basic CAD tasks: model understanding, generation, editing, and evaluation. We design test cases spanning different levels of complexity and engineering requirements. The results show that GPT-6 Astra can support multiple stages of CAD workflows, with successful outcomes in cases involving nontrivial geometric complexity. This suggests that general-purpose AI is beginning to offer practical value in almost all steps in CAD workflow. However, gaps remain between generating visually plausible geometry and producing valid, functional, and editable CAD models. Recurring errors in freeform geometry, parameter dependencies, and mating relationships between parts limit engineering usefulness. Agentic iterative refinement can improve results but does not consistently resolve critical errors; long completion times further constrain practical use.
- [812] arXiv:2610.24471 (replaced) [pdf, html, other]
-
Title: Pluralistic Alignment in Large Language Models: Objectives, Human References, and Preference RepresentationsSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Pluralistic alignment aims to represent meaningful variation in human preferences rather than collapse heterogeneous judgments into a single target. Yet existing work often mixes three distinct questions: what pluralistic behavior a model should exhibit, whose preferences define the reference, and what evidence is sufficient to demonstrate success. We organize the literature around a two-layer taxonomy. Layer 1 distinguishes representing a range of human-supported viewpoints (Overton Pluralism), adapting faithfully to a specified perspective or preference condition (Steerable Pluralism), and matching the response distribution observed in a target population (Distributional Pluralism). Layer 2 distinguishes aggregate populations, countries and regions, and subgroups as human references. Preference representation, implementation mechanism, modality, and evidence quality are treated as cross-cutting design choices rather than additional forms of pluralism. We synthesize a protocol-bounded corpus of 70 works on pluralistic alignment and neighboring areas, using the objective-by-human-reference matrix to map existing approaches and identify underexplored combinations. The coded literature concentrates on Steerable Pluralism and Distributional Pluralism, while eligible Overton Pluralism work appears only at the aggregate-population level under our stated protocol. We then connect each objective to the evidence needed to support it, including perspective validity, conditional fidelity, distributional fidelity, support preservation, and uncertainty in subgroup estimates. Together, the taxonomy and evaluation framework clarify what current methods establish, where their claims exceed their evidence, and how future systems can represent disagreement without mistaking variation for faithful representation.
- [813] arXiv:2610.26002 (replaced) [pdf, html, other]
-
Title: TimesJev: Thinking Fast on Time Series with System 1 SufficiencySubjects: Artificial Intelligence (cs.AI)
Large language models (LLMs) have recently shown strong potential for time series reasoning, while models that reason over time series are usually with long chains of thought. However, this process often consumes thousands of tokens, directly leading to substantial latency during inference. In contrast, System 1 models such as Jev make decisions instantly and can naturally be an alternative, but raw time series are not compatible with the embedding space of these models, thus hindering the reasoning performance, especially when input series contain thousands of numeric points. In this paper, we make a very early attempt to propose a time series System 1 model, namely TimesJev. Specifically, we first formalize System 1 sufficiency in time series reasoning, which motivates a compressed representation for time series with key frequency patterns preserved. To construct this representation, we derive a dimension reduction objective that combines (i) a sufficiency divergence that measures the core information lost during compression, and (ii) a realizability divergence that quantifies the incompatibility of the System 1 model to the optimal prediction given this representation. We optimize this objective by learning time series compression with a gating mechanism for sufficiency, and then aligning the selected representation with the embedding space for realizability. Finally, experiments on 13 datasets demonstrate that TimesJev outperforms reasoning models by over 62% with 62.9% less inference time than their fine-tuned versions on average.
- [814] arXiv:2610.27508 (replaced) [pdf, html, other]
-
Title: KatoANN: Enabling Near-Exact Vector Search at the Speed of Lossy Scans with CPU Matrix EnginesSubjects: Databases (cs.DB); Hardware Architecture (cs.AR)
Emerging CPU matrix engines, such as Intel AMX, enable high throughput by batching multiple queries. However, they support only low-precision formats, which can significantly degrade recall (i.e., search quality). We present KatoANN, the first approximate nearest neighbor (ANN) search system that achieves both high throughput and high recall simultaneously on CPU matrix engines. Specifically, KatoANN represents each vector using a 16-bit code, splits it into high- and low-byte planes, and computes near-exact distances using 8-bit matrix multiplications that the matrix engine supports. To reduce extra multiplication overhead, it ranks all candidates by their high bytes alone and refines only a few survivors with the full distance. KatoANN also reduces memory footprint since it does not maintain a full-precision copy of the vectors. We implement a KatoANN prototype on Intel AMX. Our experimental results show that KatoANN outperforms existing partition-based systems across the high-recall range.
- [815] arXiv:2610.28411 (replaced) [pdf, html, other]
-
Title: Brain2Vec: A Self-Supervised 3D Foundation Model for Structural Brain MRISpiros Maggioros, Guray Erus, Gareth Harman, George Aidinis, Alexander Getka, Aristeidis Sotiras, Christos DavatzikosSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Deep learning models have become essential tools in medical imaging, yet training them end-to-end for diverse tasks remains challenging due to limited labeled data and computational constraints. To address this, we introduce Brain2Vec, a foundation model pre-trained on 74,425 T1-weighted Magnetic Resonance Imaging (MRI) scans pooled from multiple studies. Brain2Vec can serve both as a weight initialization backbone for fine-tuning downstream tasks and as a standalone feature extractor. We evaluated Brain2Vec in three tasks: downstream predictive performance, convergence speed, and embedding quality. Using Brain2Vec's pretrained weights as initialization and fine-tuning on labeled data for a brain age prediction task, results demonstrate improved final accuracy and faster convergence compared to random initialization. Additionally, its frozen embeddings outperform anatomically defined volumetric features on four of five downstream classification tasks and perform competitively with a state-of-the-art foundation model pre-trained on substantially more data. On the independent BrainFMBench benchmark, Brain2Vec ranked first among six pretrained brain MRI foundation models on 9 of 15 sex, age, and BMI prediction tasks across five cohorts. Lastly, we show that Brain2Vec encodes information primarily from a spatially localized and neuroanatomically meaningful set of brain regions, and that manipulating a subset of latent features produces monotonic, region-specific effects on the reconstructed anatomy. Brain2Vec code and trained models are available as an open-source package on this https URL. Users can apply Brain2Vec locally by installing the package, or directly through the NiChart cloud platform this https URL, where Brain2Vec features can be derived without requiring local installation or specialized infrastructure.
- [816] arXiv:2610.30204 (replaced) [pdf, html, other]
-
Title: Triple Products of Correlated Multivariate Hermite PolynomialsComments: 25 pages, 4 figures, 1 supplementary PDF with data tablesSubjects: Numerical Analysis (math.NA); Methodology (stat.ME)
Stochastic Galerkin schemes (SGS) are used to approximate solutions to differential equations with stochastic parameters. In particular, they express the solution with an infinite orthogonal functional basis of the parameter space and derive a deterministic surrogate system of PDEs whose solution approximates that of the original stochastic problem. When the parameters are correlated, the infinite series representation of the solution does not provably converge for all joint distributions; further, classic decorrelating transformations (e.g., Rosenblatt, Nataf) often introduce known numerical instabilities. Rahman (2017) established convergence in probability and distribution for correlated Gaussian parameters when the basis functionals are multivariate Hermite polynomials (MHPs), where these polynomials are only weakly orthogonal with respect to the underlying joint density. Due to weak orthogonality, the $N$th moments of MHPs that emerge during Galerkin projection do not trivially decompose into products of one-dimensional expectations via tensorization. Namely, an $N$th-order moment has no closed-form expression for $N > 2$ and consequently is estimated numerically. Because the number of such moments grows combinatorially with parameter dimension and polynomial degree, approximating each one individually via sampling or quadrature can quickly become cost-prohibitive. In this paper, we introduce and prove several foundational properties of MHPs. These results culminate in an explicit formula for evaluating the $N = 3$ moments (triple products) directly without numerical estimation.
- [817] arXiv:2610.30499 (replaced) [pdf, html, other]
-
Title: Can LLMs Improve Their Own Harness? A Controlled Study on ARC-AGI-3Comments: 32 pages, 5 figures. Includes appendices and a supplementary code/data archive. v2: revised version with additional analysesSubjects: Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
A generally intelligent system acquires new skills efficiently without requiring humans to redesign its software scaffold for every new task. Yet language models rarely act alone. A harness is a symbolic program that orchestrates calls to the model, handling perception, memory, tool use, recovery, and inference allocation on the model's behalf. ARC-AGI-3 is an interactive benchmark in which agents must explore unfamiliar environments, figure out their goals on the fly, and keep learning as they go. On it, human-steered systems already report ceiling scores across the public games, so we know effective designs exist. We ask whether frontier models can build and improve such a harness on their own. Through a fixed coding bootstrap, each model constructs a complete task harness, acts through that harness on five development games, and revises the entire codebase twice. Every version is evaluated on twenty games that never influence adaptation. We measure task performance with Relative Human Action Efficiency (RHAE) and record dollar-denominated inference cost. Across 27 lineages, harness performance ranges from zero to near-ceiling scores. Two revisions raise mean held-out RHAE by only 2.80 points (95% interval [-1.44,7.35]). Development gains overstate held-out improvement, and reductions in inference cost can come at the expense of task performance. No revision retains its parent, yet a post hoc rule that selects among existing snapshots by development score raises the held-out gain to 4.53 points ([0.81,8.68]), separating the generation of revisions from the decision to deploy them. Human-steered systems demonstrate further headroom. These findings reveal a gap between autonomous harness construction and reliable adaptation under this two-revision protocol: even strong task performance does not ensure that subsequent revisions generalize or preserve existing capabilities.
- [818] arXiv:2610.30969 (replaced) [pdf, html, other]
-
Title: IORA: Irregular Object Immobilization under Restricted ActuationComments: Preprint, Project Page: this https URLSubjects: Robotics (cs.RO)
Stable grasping under task-specific disturbances is important for robotic manipulation. However, most dexterous grasp synthesis methods are developed for highly actuated hands in free space, and low-DoF hands may not be able to realize the contact topologies required by these grasps. Meanwhile, support surfaces can provide useful contact forces, but are often treated only as geometric constraints. We present a support-aware grasp synthesis framework for object immobilization with low-DoF coupled hands. Given irregular object geometry, our method represents both hand and object local surface regions as graphs and searches for kinematically feasible matchings between them. We then refine the wrist pose and motor state for geometric contact feasibility and validates the refined grasps through physical rollouts. Starting from the validated grasp pose, an optional task-guided local search explores nearby grasp configurations by varying the wrist pose and motor state, and selects one according to the specified task condition. Experiments on diverse objects show that our method generates various grasp configurations that achieve high grasp success and strong disturbance resistance. Real-world experiments further demonstrate physically executable grasps which enable object immobilization in cooking scenario.
- [819] arXiv:2610.31552 (replaced) [pdf, html, other]
-
Title: CAMGen: A Core Change-Aware Multi-View Framework for Commit Message GenerationComments: 32 pages, 5 figures, 8 tables, Manuscript submitted to a Journal (2026)Subjects: Software Engineering (cs.SE)
Commit messages help developers understand code changes and support software maintenance. Recent approaches use repository context to improve commit message generation (CMG) with large language models (LLMs). However, existing CMG methods provide limited support for understanding how local edits affect functionality jointly implemented by multiple code entities. This functional understanding is essential for generating commit messages that explain the main change and its impact. To address this problem, we propose CAMGen, a core change-aware multi-view commit message generation framework that incorporates functional context into generation. CAMGen identifies core entities that represent the primary change intent. It then forms candidate functional units through repository-wide semantic clustering and uses those containing the core entities to construct a Functional Semantic View. This view supplies semantic context for interpreting the affected functionality. CAMGen also uses context provided by the Structural Dependency View and the Historical Exemplar View. The context from all three views is then provided to the LLM together with the code diff to guide CMG. Experimental results across three programming languages show that CAMGen outperforms representative CMG baselines, including LLM-based approaches, on most automatic evaluation metrics. Human evaluation further shows that CAMGen generates more complete commit messages with clearer explanations of change rationale and impact. These results demonstrate the value of functional context and complementary repository information for CMG.
- [820] arXiv:2610.31840 (replaced) [pdf, html, other]
-
Title: SPECMEM: Accelerating Agentic Tool Calling via Live Memory ManagementComments: Accepted to Findings of EACL 2027Subjects: Computation and Language (cs.CL)
Tool-calling agents are mostly optimized for accuracy; in deployment, however, their growing cost is latency: every tool call is decoded autoregressively. Speculative decoding accelerates this without changing the output, and retrieved past calls make strong drafts under real use, where each user's behavior tends to repeat. Yet existing systems keep no cross-session memory, or build the draft store once and never update it. We propose SpecMem, a tool-call memory kept alive and managed throughout deployment: verified calls are written back online into bounded per-user stores with LRU eviction, and a side-effect gate defers irreversible calls until verification confirms. On multi-session, multi-user query streams served by three open-weight models, a frozen store steadily loses acceptance while SpecMem holds it: a 27% gap in mean accepted tokens (+11% to +34% across benchmarks, tracking how much novel work arrives after the store is built) and up to 1.74x wall-clock speedup. Ablations attribute the gain almost entirely to keeping the store live; local per-user scoping adds no acceptance but closes a measured cross-user leakage channel, and gating on declared side effects, rather than on retrieval confidence, is what eliminates irreversible mistakes.
- [821] arXiv:2610.32664 (replaced) [pdf, html, other]
-
Title: Zoom Where the Answer Looks: Cross-Resolution Self-Distillation for MLLMs under Visual-Token BudgetsSubjects: Computer Vision and Pattern Recognition (cs.CV)
Under the small per-view visual-token budgets of efficient deployment, the fine text of documents, charts and infographics becomes illegible to multimodal large language models (MLLMs). Zooming in recovers it, but the region must be chosen before the model answers. Existing methods decode an answer or a box first, at the cost of generation, or read the attention available before answering, which localizes the evidence only in top layers whose depth varies with model size. Learned region proposers instead distill the attention of the model's answers, but on the same budgeted view. Yet this attention is precise only on a view the model can read: on Qwen3-VL-4B, a window around its peak covers 83% of the evidence on a high-resolution view but 54% on the budgeted view. Since answers and high-resolution views are available during training, we introduce LookZoom, a zooming method that trains a dedicated look token by cross-resolution self-distillation of the answer's attention: with the MLLM frozen, the token learns to attend on the budgeted view, before any answer, where the model's own correct answers attend on a high-resolution view. The region around its attention peak is the window to zoom into, and reinforcement learning then trains the MLLM to answer from both views. With 98 visual tokens per view on five text-rich benchmarks, LookZoom surpasses attention-guided zooming by 20-24 points and grounding-based zooming and learned crop routing by 12-15 points on Qwen3-VL at three scales, and it outperforms learned region proposers at every budget from 98 to 392 tokens. On Qwen3-VL, the look token alone, with the MLLM frozen, already surpasses every baseline that leaves the language model unchanged, and the recipe carries over to InternVL3.5 and Qwen2.5-VL.
- [822] arXiv:2610.32907 (replaced) [pdf, html, other]
-
Title: A Highly Efficient Depth-Averaged Material Point Method for Geophysical Mass FlowsComments: v2: Added subsection on computational performance (otherwise results unchanged); version submitted to Computers & GeosciencesSubjects: Computational Engineering, Finance, and Science (cs.CE)
Probabilistic hazard assessment of geophysical mass flows requires large ensemble simulations and thus fast and robust numerical models. We present an open-source depth-averaged material point method formulated in global Cartesian coordinates, using Lagrangian particles to track the flow margin. A separate basal friction step based on the Coulomb, Voellmy, or $\mu(I)$-rheology brings the flow to rest without artificial stopping criteria. Our model reproduces analytical solutions for a sliding block and dam-break problems and matches the deposit of a laboratory granular column collapse. A back-calculation of the 2023 Brienz/Brinzauls rockslide (Switzerland) at 1m resolution (190'000 particles) runs faster than real time on a single CPU core. Calibrated to the observed runout, the three friction laws yield very similar results, but Coulomb friction shows a slightly wider affected area than Voellmy and $\mu(I)$. A probabilistic 1632-simulation ensemble at 5m resolution is computed in 76s. For a single simulation at 1m resolution, the model is 69 times faster than RAMMS::AVALANCHE and 16 times faster than AvaFrame com1DFA; for 1000 simulations at 5m resolution, it is 540 and 67 times faster, respectively, making catchment-style ensemble simulations feasible on a desktop computer.
- [823] arXiv:2610.33253 (replaced) [pdf, html, other]
-
Title: Operationalizing Governed Agentic Workflows in Regulated Finance: A Customer Zero Case Study using Zafin AIOSShahir Daya, Rob Dunlap, Branavan Selvasingham, Alexandre Boulenger, Ahmed Abdelmoneim, Salvatore VellaComments: 19 pagesSubjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI)
Regulated industries such as finance and healthcare face critical governance challenges that hamper the adoption of AI-assisted software development and hinder the transition from experimental pilots to production work. This paper presents Zafin AIOS, an end-to-end orchestration platform and control plane designed to deploy AI within regulated industries such as finance. We provide the architectural principles and an empirical comparison between AI-assisted and human-led workloads. To validate this framework empirically, we conduct a case study analysis of Zafin's internal enterprise workflows, which we refer to as Customer Zero. This paper contributes a production case study of AI-assisted software engineering in regulated industries. We evaluate the results of using Zafin AIOS with two different production codebases, one for the pricing execution platform and the other for the client deal negotiation platform. Our contribution is a customer case study for regulated industries with production codebases and governance, which includes proof of work. The results demonstrate that strictly governed agentic orchestration increases enterprise capacity without compromising control. In the pricing execution platform workflows, the platform provided complete observability and reduced effort by 38\%. In the client deal negotiation platform, it reduced reporting work by 76\% while maintaining all predefined data privacy, role-based access, and audit guardrails. This paper bridges the gap between AI-assisted software engineering theory and the operational reality of compliance. Closing this gap is important to help enterprises take advantage of AI technologies. These results are descriptive of the benefits of using governed workflows in regulated industries.
- [824] arXiv:2610.36145 (replaced) [pdf, html, other]
-
Title: They Will See Johnny's Message: A Measurement Study of Post-Quantum Cryptography Across Email Delivery PathsComments: 12 pagesSubjects: Cryptography and Security (cs.CR)
Harvest-now, decrypt-later (HNDL) attacks make the transition to post-quantum cryptography (PQC) urgent for email. This paper presents a systematic measurement that jointly examines PQC deployment across email service stages, negotiation by communicating peers, and usable message encryption. Endpoint-level support alone does not establish PQC readiness across email workflows because submission, relay, and access use independently negotiated Transport Layer Security (TLS) connections. We therefore combine three snapshots of relay, submission, access, and webmail services associated with the Tranco Top~1M domains with controlled tests of mail clients, sending servers, and OpenPGP and S/MIME implementations. Our measurements reveal a transition concentrated among a few providers, with available capability not consistently translating into actual use. Fewer than half of the domains in each service panel support PQC, and Google alone accounts for approximately 87\% of hybrid-capable relay domains. Only two of eight tested sending providers negotiate PQC with our receiver, although five accept it inbound. When TLS groups are restricted to PQC, four of the six incompatible senders fall back to plaintext where unencrypted SMTP remains permitted. At the message layer, limited native key generation, key-format incompatibilities, and backend integration failures further constrain usable PQC encryption.
- [825] arXiv:2610.39180 (replaced) [pdf, other]
-
Title: EmbodiExp: Compositional Experience Transfer for Embodied AgentsSubjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Experience from past tasks can help embodied agents perform new tasks, but it may not remain applicable or effective when conditions change. Tasks with different overall goals often share subgoals, providing a basis for cross-task experience transfer. We present EmbodiExp, a framework for constructing outcome-grounded subgoal experiences and composing them in new tasks. Each experience describes how to pursue a subgoal, when the guidance applies, and what outcome must be verified for completion. Verified outcomes from applying one experience can satisfy the conditions for applying another, allowing experiences from different past tasks to be composed online. By examining the actions and observed outcomes at the subgoal level, EmbodiExp retains experience from successful subgoals and uses explicit action failures to identify actions to avoid under specific conditions. Experiments on EmbodiedBench and VIGIL assess cross-task transfer with experiences kept fixed throughout held-out evaluation. EmbodiExp achieves success rates of 68.0% on EB-Habitat and 68.3% on EB-ALFRED, compared with 50.0% and 54.7% for the agent without stored experience.
- [826] arXiv:2610.40143 (replaced) [pdf, html, other]
-
Title: Mode Dynamics: A Framework for Neural Network LearningComments: 37 pages, 18 figures, 8 tablesSubjects: Machine Learning (cs.LG)
We introduce Mode Dynamics, a framework that connects coordinated parameter motion during neural network learning to changes in model outputs. Starting from a recorded parameter trajectory, we decompose its updates into collective modes and map each mode to its state-dependent output effect. The framework organizes this link through three mechanisms. Decomposition identifies collective modes in parameter motion and their time-varying amplitudes. Reconstruction tests whether the effects induced by those modes recover local and accumulated output change. Concentration quantifies how behavioral contribution is distributed across modes. Controlled scalar regression and next-token experiments directly test these mechanisms. They show that small sets of leading modes reconstruct behavior, induced effects depend on the training path, and modal complexity varies with task and data. In a controlled induction task, suppressing the leading modes during training impairs held-out accuracy, whereas matched controls do not. We then evaluate the framework on held-out image and language-model outputs using CIFAR-100 autoencoding and a public Pythia-70M pretraining trajectory. Across these settings, the results connect compact modal subspaces of recorded parameter motion to evolving behavior.
- [827] arXiv:2610.40304 (replaced) [pdf, html, other]
-
Title: Causal-prior-constrained deep learning for rapid-intensification warning of western North Pacific tropical cyclonesComments: 34 pages, including 15 pages of supplementary information; 5 main figures and 6 supplementary figuresSubjects: Machine Learning (cs.LG); Atmospheric and Oceanic Physics (physics.ao-ph)
Rapid intensification (RI) is one of the most dangerous and unpredictable phases of tropical cyclones, and its environmental and structural drivers are difficult to identify and separate in observations. Here we introduce CausalVisionRI, a causal-prior-constrained deep-learning system for 24-h RI warning in the western North Pacific. It integrates infrared remote sensing imagery, atmospheric and oceanic fields, and an observationally discovered causal graph into a unified model. The graph and eight sign-verified effects on 24-h intensity change provide an auditable causal prior for training. Discovered from 522 cyclones in 2000-2017, it identifies causal pathways linking atmospheric and oceanic conditions, the cyclone's dynamical and structural state, and its 24-h intensity change. On a forward holdout covering 2020-2024 and 3,004 records, the five-member ensemble achieves a Brier skill score (BSS) of 0.259 relative to a fixed development climatology, significantly exceeding four reference systems; the three LDA baselines reach at most 0.105. It also attains higher detection rates than a published knowledge-block scheme and a reconstructed contrastive-learning scheme at matched false-detection rates under our USA 30-kt standard and the published RITCF standard. Ablations show that the graph pathway adds 0.032 BSS across the holdout and 0.063 over 2020-2021, indicating that the gain concentrates in the time period adjacent to the estimation window. A deep-learning model constrained by observationally discovered causal structure thus offers a new route to physically informed RI warning.
- [828] arXiv:2610.41354 (replaced) [pdf, html, other]
-
Title: OneSteer: Unlocking Steerable General Humanoid Motion Tracking via One-Step Diffusion Whole-Body ControlYifei Huang, Yazhao Yang, Li Weijia, Tianle Shi, Ji Jin, Zhihui Cai, Zihao Liu, Jianzhuang Zhao, Songlin Dong, Xinyuan Chang, Yihong Gong, Dian Shen, Zhiheng MaSubjects: Robotics (cs.RO)
Deployment objectives often specify desired outcomes without providing revised trajectories, while parts of the original reference must still be followed. We present OneSteer, a one-step diffusion whole-body controller that combines general motion tracking with selective test-time steering. To accommodate new objectives, differentiable task costs guide optimization of the generator's input noise with frozen weights, while body-group reference masks relax selected tracking requirements. Predicted state-action plans provide a shared interface for task optimization and feedback-conditioned execution. For nominal tracking, the model trained from scratch achieves the lowest local tracking error among the evaluated baselines on successful held-out AMASS and BONES clips. For steering, one frozen distilled controller adapts across diverse deployment objectives in simulation, raising the clean passage rate under a 1.25 m bar from 2% to 90%. On a physical Unitree G1, OneSteer completes all 27 nominal tracking trials across nine diverse whole-body motion sequences without falling. With a steering cost derived from a signed distance field built using only onboard LiDAR, it also passes under a 1.1 m low bar in four of five attempts and steps over a 12 cm hurdle without falling in all three attempts. Beyond control, prediction variation in the model trained from scratch ranks likely tracking failures without physics rollouts. This scoring method is more than 10x faster than simulation-based scoring with five rollouts per clip and guides targeted fine-tuning. Project page: this https URL
- [829] arXiv:2610.41525 (replaced) [pdf, html, other]
-
Title: MemX: Memory-Aware Automatic Parallelism for Cross-Vendor Heterogeneous GPU TrainingComments: Code: this https URLSubjects: Distributed, Parallel, and Cluster Computing (cs.DC)
Cross-vendor heterogeneous GPU clusters are increasingly used for large-scale model training. However, their GPUs differ substantially in compute power, memory capacity, and interconnect characteristics. Existing automatic parallelism training systems usually minimize iteration time while treating memory as a hard constraint, almost failing in resolving the compute--memory asymmetry of cross-vendor clusters regardless correlation between memory and throughput. In this paper, we present MemX, a new memory-aware automatic parallelism system that jointly optimizes iteration time, peak memory usage, and cross-GPU memory balance. MemX consists of a weighted-product objective function, heterogeneous recomputation strategies, a minimal-set profiler, interpolation-based data augmentation, a dual-output XGBoost predictor, and a hash-signature prediction cache, in order to accelerate search of large-scale auto-parallel best configurations at lower cost. We implement MemX on Megatron-LM and evaluate it against four automatic parallelism baselines on two heterogeneous clusters. MemX achieves the lowest prediction Mean Absolute Percentage Error (MAPE) of 12.17% and is the only method that identifies the real Top-1 configuration under both time-based and memory-balance rankings. In terms of search cost and search space, MemX searches 31,240 configurations in 26.5 seconds, while an end-to-end profiling baseline is extrapolated to require about 17 days at the same scale. The simulation study that extends the measured 20-GPU data to 40-GPU and 60-GPU clusters further illustrates the effectiveness of MemX at larger scales.
- [830] arXiv:2610.42121 (replaced) [pdf, html, other]
-
Title: Where ADS Crashes Differ: A Harmonized Comparison Against Human BaselinesSubjects: Computers and Society (cs.CY)
Crash-rate comparisons between automated driving systems (ADS) and human drivers require an exposure denominator that only a few operators publish. We instead ask a composition question: given a crash, does the mix of crash types differ between ADS and human-driven vehicles? The question needs no mileage and can therefore cover every ADS operator. We assemble the NHTSA Standing General Order (SGO) ADS incident files across both file generations into 2,450 distinct in-scope incidents (2021-2026), of which 828 fall within the 2021-2024 window covered by the Crash Report Sampling System (CRSS), the national police-reported human comparator. A schema-constrained large language model (LLM) codes each SGO narrative into the CRSS crash-type taxonomy. Because SGO compels reporting far below the police-reporting threshold, we bracket reportability with an objective severity floor and a reasonable-person rubric, and claim a difference only where it survives both. Two composition differences hold under every reportability definition, tow definition and weighting of the human side, and under both definitions of driverless operation: turn-across-path crashes are under-represented among ADS crashes (0.23-0.46x the CRSS share under the primary driverless definition) and so are single-vehicle crashes (0.19-0.56x). In severity-floor rear-end crashes the ADS vehicle is the striking party at a quasi-induced-exposure ratio of 0.121 (95% CI [0.05, 0.23]) against a human baseline of 1.000. We also document three pitfalls in the federal data: an operator-specific coding of driverless operation, a tow-away criterion confounded by fleet recovery practice, and an occupancy confound that accounts for about 42% of the apparent ADS-human severity gap.
- [831] arXiv:2611.00406 (replaced) [pdf, html, other]
-
Title: TAP-Out: Tracking Any Point in 360° via OutpaintingComments: Accepted at BMVC 2026. 22 pages, including supplementary material. Project page: this https URL Code: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV)
Tracking points that leave the camera's field of view (FoV) requires reasoning about unobserved scene dynamics, a fundamentally different problem from tracking visible correspondences. We present TAP-Out, an offline method that makes this out-of-view (OOV) reasoning explicit by leveraging generative video outpainting as a world-model prior. Given a narrow-FoV input video, TAP-Out outpaints a complete 360° equirectangular video, locates target objects within the panoramic scene via video object segmentation, and synthesises object-centred perspective crops on which standard in-frame trackers operate. This decoupled design composes off-the-shelf foundation models without requiring task-specific architectural modifications. On the TAPVid360 benchmark, TAP-Out achieves state-of-the-art performance, improving OOV tracking accuracy by approximately 3x over the best prior method. We demonstrate its flexibility on out-of-frame object segmentation and detection and we provide detailed diagnostic analyses across motion regimes, static versus dynamic content, and hallucination quality, characterising when and why the approach succeeds or fails. Code is publicly available at this https URL
- [832] arXiv:2611.00430 (replaced) [pdf, html, other]
-
Title: Real-VAS: a Real-World Video Amodal Segmentation DatasetComments: 34 pages, including supplementary material. Project page: this https URL Code: this https URL Dataset: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV)
Amodal video object segmentation is fundamentally limited by the absence of datasets that combine real-world complexity with precise ground-truth annotations. To address this, we present Real Video Amodal Segmentation (Real-VAS), a new large-scale, zero-shot evaluation dataset built from composited real-image video. We introduce a novel data generation pipeline that enables the creation of programmatically consistent amodal ground truth for real-world footage, in which the object mask, occlusion boundary, and composited image are all derived from a single unoccluded source, without relying on human estimation or expensive 3D reconstruction. Our dataset is structured into two challenging scenarios: dynamic Occlusion, created by compositing two clips of moving objects, and a unique Container category featuring complex, physically constrained interactions. These container scenarios--On Surface Containment, Articulated Containment, and Mobile Containment--allow us to generate precise ground truth by simulating an object's motion based on its container's tracked transformation. As a result, Real-VAS provides a diverse and challenging benchmark for evaluating amodal segmentation models on realistic video with the precision of synthetic data. The dataset and generation code are publicly available at this https URL
- [833] arXiv:2611.00482 (replaced) [pdf, html, other]
-
Title: A Straight Line Is Not a Camera Pose: Auditing Short-Clip Pose BenchmarksComments: 39 pages, 9 figuresSubjects: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
Many vision benchmarks align a prediction to the ground truth before scoring it. If the alignment minimises the error, an input-free template scores the truth's distance from everything the alignment can reach, so the ground truth alone fixes how well a template can score, before any model runs. Camera pose is our central case. The absolute trajectory error (ATE), the standard camera-motion measure, aligns camera positions by a similarity transform. On the common Sintel protocol, an image-free constant-velocity line has a lower mean ATE than 10 of 11 recent models and 85 of 110 published methods, and on Lightspeed than 5 of 9 models and 10 of 11 published methods. The straightness of the true paths separates these two from TUM, ScanNet, DyCheck and ADT, where every model beats the line. A residual signal-to-noise criterion with the alignment's first-order refit predicts when straightening a prediction lowers ATE in all but one of 4010 model-clip evaluations, including all 504 on three benchmarks registered in advance. The alignment also ignores where cameras look: rotating a trajectory by 30 degrees against its orientations leaves ATE unchanged yet worsens the static-point reprojection of every model on Sintel, TUM and ScanNet. Fitting the alignment's rotation to the predicted orientations gives pose-anchored ATE (PA-ATE), a one-function change at about ATE's cost (37 vs. 34 microseconds per clip). It is never below ATE, ranks every image-free reference below every model on Sintel, tracks the reprojection error under controlled rotation and straightening, largely preserves the models' order, and exposes Lightspeed ground-truth orientations that the images contradict. Beyond poses, templates beat some published methods on KITTI video depth, SSP-3D body shape and NoW faces, and in video depth this distance sets KITTI apart from Sintel and Bonn.
- [834] arXiv:2611.00782 (replaced) [pdf, html, other]
-
Title: One Offset, Two Directions: Baseline Adjustment via Subgroup OffsetsSubjects: Artificial Intelligence (cs.AI)
Group Relative Policy Optimization (GRPO) estimates advantages against the within-prompt mean reward, a baseline shared across the group. In practice, rollouts within a prompt are not always homogeneous, differing in their solution strategies and reasoning trajectories, along which they may group into subgroups. We further find that these heterogeneous subgroups may exhibit distinct reward distributions. The global baseline, however, collapses them onto a single credit axis, where the credit reflects neither a rollout's standing within its subgroup nor the subgroup's standing relative to the rest. We therefore study how to adjust the baseline when the rollouts of a prompt fall into heterogeneous subgroups. We define the \emph{subgroup offset} as the deviation of a subgroup mean from the global mean, and shift the baseline along it by a single signed coefficient $\lambda$. The sign of $\lambda$ yields two opposing directions: \emph{Alignment} ($\lambda{>}0$) adds the offset to the baseline, drawing credit toward a rollout's relative position within its subgroup; \emph{Contrast} ($\lambda{<}0$) subtracts it, drawing a rollout's credit toward the relative quality of the subgroup it belongs to. The adjustment acts on the baseline alone, leaving the reward signal and the optimization objective unchanged. In this work, we induce subgroups from strategies, grouping rollouts that share a strategy into the same subgroup. We evaluate both directions on Qwen3-8B-Base and Olmo3-7B-Base across four competition-level math benchmarks, against a $\lambda{=}0$ anchor that shares the same strategy-conditioned sampling and differs only in the baseline. Against this anchor each direction yields modest but consistent gains in multi-sample coverage, with Contrast the stronger of the two.
- [835] arXiv:2611.00793 (replaced) [pdf, html, other]
-
Title: A Second-Order Stochastic Dynamic Mode Decomposition by Richardson ExtrapolationSubjects: Numerical Analysis (math.NA)
Stochastic dynamic mode decomposition (SDMD) approximates the Koopman semigroup of an Itô diffusion by a Galerkin generator matrix on a dictionary of observables, computed from samples of the invariant law and generator evaluations, and advances observables with a first-order stochastic Taylor step (SDMD-1). The second-order Taylor step (SDMD-2) also needs a third matrix, a dictionary with four derivatives, and a fourth-order differential operator evaluated at every sample. Richardson extrapolation of the SDMD-1 step (SDMD-R) is second-order accurate with the data, the matrices, and the regularity of SDMD-1. The error bound, for exact and empirical matrices, consists of the Galerkin error of the dictionary, a time-stepping term, and a data term of order $m^{-1/2}$ in the sample size $m$. Relative to the Galerkin semigroup, the limit of the schemes as the time step $\dt\to0$, the time-stepping error is $O(\dt^2)$ for SDMD-R and $O(\Theta_N\dt+\dt^2)$ for SDMD-2, with $\Theta_N$ the norm of the second-order Galerkin residual $\Proj_N\cA(I-\Proj_N)\cA|_{\cV_N}$, and the first-order term is attained. For reversible periodic diffusions and trigonometric dictionaries with cutoff $K$, the step restriction $\dt\lesssim K^{-2}$ makes all three schemes contractions, and the error against the Koopman semigroup on $H^\ell$ observables is $O(K^{-(\ell-2)}+\dt^p)$ plus a data term of order $K^{d+2}(m^{-1}\log K)^{1/2}$ in dimension $d$, with $p=1$ for SDMD-1 and $p=2$ for SDMD-R and SDMD-2. In experiments with trigonometric dictionaries the errors of SDMD-R and SDMD-2 differ by at most ten percent. For a spline dictionary with two derivatives, SDMD-R remains second-order accurate, whereas SDMD-2 with pointwise evaluations is first-order accurate.
- [836] arXiv:2611.01514 (replaced) [pdf, html, other]
-
Title: Federated Vector Retrieval across Heterogeneous VectorDBsComments: Accepted for publication at VLDB 2027. The manuscript contains 13 figuresSubjects: Databases (cs.DB); Information Retrieval (cs.IR)
Vector databases are now core infrastructure for AI retrieval systems, yet many real deployments must serve queries over multiple independently constructed vectorDBs built with different embedding models. This creates a systems problem of federated retrieval over heterogeneous vectorDBs: vectors and similarity scores are not directly comparable across embedding spaces, while fully resolving this incompatibility through corpus-wide re-embedding and re-indexing is often operationally prohibitive. Existing methods therefore face a difficult trade-off between retrieval quality, construction cost, and query throughput. We present HVMap, a lightweight compatibility layer for federated retrieval across heterogeneous vectorDBs. Instead of rebuilding databases into a unified space, HVMap learns bidirectional mappings between embedding spaces using a fuzzy contrastive loss, enabling calibrated cross-database scoring without corpus migration or index reconstruction. HVMap is system-agnostic and further scales to multi-vectorDB settings while preserving ranking fidelity through a hub-based architecture. We evaluate HVMap in open-domain, cross-domain, and scale-out settings, and also construct a new benchmark for heterogeneous cross-domain federation. Across these settings, HVMap consistently achieves a better quality--cost--throughput trade-off than baselines, achieving up to 3.59X higher nDCG@10, up to 2.01X higher query throughput, and construction time as low as 0.3% of baseline methods. These results show that HVMap makes heterogeneous vectorDB federation practical at scale.
- [837] arXiv:2611.01929 (replaced) [pdf, html, other]
-
Title: Watch Your Step: Adam's Margin Maximization Bias Depends on the Learning Rate ScheduleSubjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
We study the implicit bias of Adam and its similarity to sign gradient descent on homogeneous models in binary classification tasks. Perhaps surprisingly, we uncover that these depend on the learning rate schedule. While previous works studied strongly decaying learning rate schedules $\eta(t)=\mathrm o \left(t^{1/L-1}\right)$, where $L$ is the homogeneity degree, and showed $\ell_\infty$-margin-maximization biases, we focus on the more common complementary regime, including constant and weakly decaying schedules. For a polynomial schedule $t^{\alpha}$ and homogeneity degree $L$, we show that when $\alpha > \frac 1L - \frac12$, Adam has no trajectory-independent margin-maximization bias for deep linear networks (in the KKT sense): first, there is no non-trivial, locally Lipschitz, convex implicit regularizer in this setting, including any squared norm. Second, on width-$1$ networks, the result holds without requiring convexity, implying that any implicit regularizer for wide networks must be trivial when restricted to a width-$1$ network. Nevertheless, we positively characterize the implicit bias in this regime as maximizing a trajectory-dependent weighted-$\ell_2$ margin for any $L$-homogeneous model. The fundamental mechanism behind our results is convergence of Adam's preconditioner to a trajectory-dependent vector, which also allows showing that the limiting parameter direction depends on initialization. In the intermediate regime $\frac 1L - 1 \leq \alpha \leq \frac1L - \frac12$ we prove an $\ell_\infty$ bias, conditional on gradient directional convergence to a full-support vector. Of independent interest, we establish a loss decay rate for full batch Adam on general homogeneous models, whose leading order depends only on the second-moment parameter.
- [838] arXiv:2611.03301 (replaced) [pdf, html, other]
-
Title: An Axiomatic Framework for Reinforcement Learning in Open Multi-Agent SystemsComments: 43 pages, Published at NeurIPS 2026Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
Open multi-agent systems, where only a subset of agents are controllable, are central to real-world domains such as smart grids, yet lack a principled foundation for reinforcement learning algorithm design. We study this setting through n-agent ad hoc teamwork (NAHT), a recent framework for learning in open multi-agent environments. We propose an axiomatic framework that derives value function learning from the Shapley axioms (Efficiency, Symmetry and Linearity), treating them as structural constraints rather than heuristic design choices. This formulation builds on a novel NAHT decomposition, which characterises the fundamental difference between open and fully controllable multi-agent systems, and establishes a cooperative game-theoretic foundation for learning. We further show that enforcing these axioms on learning individual value functions recovers the Shapley value for Dec-POMDPs, providing a principled bridge between reinforcement learning and classical cooperative game theory. Building on this framework, we introduce Shapley Machines and Banzhaf Machines, two generic algorithmic templates that can be instantiated with standard reinforcement learning methods. When applied to strong base algorithms such as IPPO and POAM, our methods consistently improve learning efficiency and generalisation to unseen agent types and conventions across NAHT benchmark tasks. The code of experiments is published on this https URL.
- [839] arXiv:2611.03313 (replaced) [pdf, html, other]
-
Title: Regular Separability for Bidirected VASSSubjects: Formal Languages and Automata Theory (cs.FL)
The regular separability problem asks, given two languages, whether there exists a regular language that contains one of them and is disjoint from the other. We study regular separability for bidirected VASS. We first show that the problem remains Ackermann-complete when the input consists of two languages recognised by bidrected VASS. To find a subclass of VASS in which the problem has lower complexity we define alphabetically bidirected VASS as ones where bidirectedness is also visible in the input letters. Our main result establishes that regular separability for alphabetically bidirected VASS is EXPSPACE-complete. The key to this result is that, in this setting, regular separability coincides with disjointness. As a first ingredient in this equivalence, we also show that for alphabetically bidirected Z-VASS, regular separability and disjointness are equivalent. Consequently, regular separability for these systems is in polynomial-time. In fact, it is log-space inter-reducible with non-solvability of integer linear Diophantine equation systems. Finally, we observe that the picture changes dramatically in the presence of a zero test, namely, we show that regular separability for alphabetically bidirected one-counter automata is undecidable.
- [840] arXiv:2611.03936 (replaced) [pdf, html, other]
-
Title: Geometric Refinement and Distributional Alignment for Sim2Real Roadside LiDAR DetectionComments: 8 pages. 1st place, Track 6 (UrbanTwin Sim2Real LiDAR, LUMPI), DriveX Grand Challenge, ECCV 2026 Workshop. Code: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
We describe our approach to the LUMPI track of the UrbanTwin Sim2Real LiDAR Challenge, in which a 3D object detector must be trained exclusively on synthetic data and evaluated on unseen real roadside LiDAR, while the synthetic point clouds are simultaneously scored for distributional realism. Our final entry scores 0.4793 combined, from a detection mAP of 0.1333 and a realism score of 0.9126. Our contribution is not a new architecture. Working from a PointPillars baseline, we obtained most of our gains from three sources. First, a diagnosis that small object failure in roadside LiDAR is a localisation problem rather than a detection problem, which led to a parameter free geometric refinement worth 86 percent relative on pedestrian AP. Second, an audit of our own post processing which revealed that a shipped rule was destroying valid predictions, the repair of which produced our single largest detection gain at 335 percent relative on the Van class. Third, treating the realism metrics analytically rather than as a black box, which showed that one of the four components is a closed form function of nine summary statistics. We also report an experimental methodology that we believe is the more transferable result. Because mean AP decomposes into independent per class terms, a single submission can carry several mutually controlled experiments, and a submission assembled entirely from previously measured components has a score that is known before it is sent. We predicted our final score to four decimal places prior to submission. We report negative results at equal length, and give a full account of data provenance and compliance in Section 8, including a full account of every mechanism contributing to the submitted score.
- [841] arXiv:2611.03940 (replaced) [pdf, html, other]
-
Title: Coordinate Frame Recovery, Per Class NMS and Null Return Modelling for Sim2Real V2X LiDAR DetectionComments: 8 pages. 1st place, Track 7 (UrbanTwin Sim2Real LiDAR, V2X-Real), DriveX Grand Challenge, ECCV 2026 Workshop. Code: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
We describe our approach to the V2X-Real track of the UrbanTwin Sim2Real LiDAR Challenge, in which a 3D object detector must be trained exclusively on synthetic data and evaluated on unseen real roadside LiDAR from an Ouster OS1-64 unit, while the synthetic point clouds are simultaneously scored for distributional realism. Our final entry scores 0.7457 combined, from a detection mAP of 0.4382 and a realism score of 0.9251. Three findings account for most of our result, and all three were diagnostic rather than architectural. First, a coordinate frame mismatch between our predictions and the evaluation ground truth was suppressing detection AP entirely; recovering it took our score from 0.3520 to 0.5147 in a single submission. Second, a single global non maximum suppression threshold was destroying genuine pedestrian detections, because pedestrians occur in groups and a 0.3 IoU threshold applied to a 0.7 m box suppresses distinct neighbouring people; correcting this per class raised overall mAP from 0.0751 to 0.4295. Third, approximately 31 percent of the real reference point cloud consists of null returns encoded as the literal point (0, 0, 0), a format feature entirely absent from our synthetic clouds, which pinned Earth Mover's Distance near its floor until we modelled it. Negative results are reported at equal length, and Section 9 gives a full account of data provenance and compliance. That section also accounts for every mechanism contributing to the submitted score.
- [842] arXiv:2611.04221 (replaced) [pdf, html, other]
-
Title: Mamba3R: Fixed-Capacity Selective State Memory for Long-Sequence 3D ReconstructionComments: 12 pages, 4 figures, 8 tables. The supplementary material referenced in the text is not included in this versionSubjects: Computer Vision and Pattern Recognition (cs.CV)
Streaming 3D reconstruction must maintain geometric quality over long inputs within a fixed GPU budget. A pretrained recurrent model provides bounded state, but freezing its backbone also fixes its learned state-observation interactions. We introduce Mamba3R, a jointly designed state-memory adapter that provides a trainable adaptation path without updating the backbone or task heads. A selective state-space mixer jointly transforms native state features and a fixed-capacity auxiliary memory, explicit read and write paths carry information across frames, and a decomposed feedback path regulates what the added memory feeds back. Only 1.28 million adapter parameters are trained. Under a shared periodic-reset protocol, it lowers trajectory error and improves accumulated surface and depth accuracy over recurrent baselines on public indoor and outdoor datasets and robot-recorded traversals, gaining up to 40 percent on one metric. Matched comparisons support both the choice of the mixing operator and the utility of carrying the auxiliary memory across frames. Allocated memory remains unchanged and throughput moves by at most 1.4 percent from one to two thousand frames. Replaying its recorded traversals offline, the platform's RTX 4060 averages 11.7 frames per second within 5.1 GiB with the same checkpoint, a mean throughput covering the 10 Hz sensor rate.
- [843] arXiv:2611.04240 (replaced) [pdf, html, other]
-
Title: VerfGraph: Efficient Zero-Knowledge Verifiable Graph Query Evaluation SystemComments: 14 pages, 3 figures, 7 tablesSubjects: Cryptography and Security (cs.CR)
Graph databases are widely used to manage relationship-rich data and support multi-hop traversal and shortest-path queries in applications such as financial risk control and social networks. When a client issues such queries to a database owner, the returned result may be incorrect. However, the client cannot directly verify it because the graph is typically large and private, making re-executing the query both infeasible and insecure. Verifiable and confidential graph query evaluation thus plays an important role in such applications.
We present VerfGraph, a zero-knowledge verifiable graph query evaluation system, which enables a client to verify that the query result was correctly computed over a committed graph, without revealing anything beyond the public inputs and what is implied by the query result. While adopting the expansion-centric operator model introduced by ZKGraph (ICDE'26), VerfGraph improves both proving efficiency and query expressiveness in two key ways. First, we propose new proof methods for the core operators in ZKGraph, such as graph expansion and shortest path, that require substantially fewer expensive comparisons and eliminate edge canonicalization for bidirectional edge relations. Second, we introduce dedicated proofs for count and weighted shortest-path operators, which ZKGraph does not support, broadening the class of graph queries that can be verified. We evaluate VerfGraph using the LDBC Social Network Benchmark. Compared with the state-of-the-art zero-knowledge graph query system ZKGraph on complex IC queries, VerfGraph achieves $33.9\times$-$59.7\times$ speedups in proof generation and reduces peak prover memory by a factor of $4.5\times$-$6.6\times$, while maintaining millisecond-level verification and producing sub-MB proofs. - [844] arXiv:2611.05205 (replaced) [pdf, html, other]
-
Title: TAQ-MOTR: Tracking-Aware Detection Query Formation for End-to-End Multiple Object TrackingSubjects: Computer Vision and Pattern Recognition (cs.CV)
Tracking-by-propagation (TBP) methods perform end-to-end multi-object tracking by propagating track queries across frames, jointly handling detection and identity association without heuristic post-processing while naturally modeling temporal information. However, as detection and association are tightly coupled, detection queries in TBP are expected to discover currently untracked objects and provide complementary current-frame evidence for challenging tracks whose queries and identities are retained. Existing methods largely inherit detection-query formation from single-frame detection and construct such queries before the decoder with little awareness of trajectory states. Consequently, currently untracked objects may be missed, while challenging tracked objects with reduced visibility under occlusion may lack complementary evidence, causing association failures or duplicate responses. To address these issues, we propose TAQ-MOTR, a tracking-aware detection query formation framework that employs a pre-decoder Hybrid Query Generator (HQG) to fuse current-frame image features, track memory, and learnable memory. By conditioning detection queries on trajectory states, TAQ-MOTR adapts them to \rev{the discovery of currently untracked objects and tracked-object complementation We further introduce a training-only auxiliary detection branch whose one-to-one label assignment is decoupled from that of the main tracking branch, providing additional detection supervision to strengthen the shared HQG and decoder for the discovery of currently untracked objects. TAQ-MOTR improves performance on DanceTrack, SportsMOT, and BFT, achieving 72.0 HOTA, 63.1 AssA, and 76.6 IDF1 on the DanceTrack test set. Code is available at this https URL
- [845] arXiv:2611.05546 (replaced) [pdf, html, other]
-
Title: FedSharp: Byzantine Robust Decentralized Federated Learning Under Noisy ChannelComments: 6 pages, 4 figures, submitted for possible publicationSubjects: Machine Learning (cs.LG)
Decentralized federated learning (DFL) eliminates the central aggregation server but remains vulnerable to non-independent and identically distributed (non-IID) data, Byzantine attacks, and channel noise. We show that channel noise accumulates as models propagate across nodes and becomes increasingly significant as the learning rate decays. Moreover, accumulated noise enables Byzantine attackers to exploit loss-based model selection by mimicking the average of the honest models. To address these challenges, we propose FedSharp (screened, hop-aware, averaged, ranked, proximal), a robust DFL framework. In FedSharp, each node keeps the latest models received from its 2b predecessors, which guarantees at least b honest candidates when at most b nodes are Byzantine. FedSharp screens suspicious models using their deviations from the coordinate-wise median, averages the b lowest-loss candidates to reduce the fresh channel-noise variance by a factor of b, and applies hop-aware gradient-norm regularization and proximal training to mitigate accumulated noise and non-IID drift. Experiments on a 100-node ring with 33 Byzantine nodes show that FedSharp achieves 93.5 to 94.9 percent accuracy on MNIST, compared with 29.4 to 38.1 percent for a state-of-the-art single-model-selection benchmark, and 32.8 to 48.8 percent on CIFAR-10, compared with 10.0 to 30.5 percent, across Gaussian and hidden attacks under IID and non-IID data. On MNIST, FedSharp comes within 1.3 percentage points of the benchmark operating without channel noise, and exceeds it under the hidden attack.
- [846] arXiv:2611.05798 (replaced) [pdf, html, other]
-
Title: Function-Preserving Audits of Visual Weight ReadersSubjects: Artificial Intelligence (cs.AI)
A neural field can encode the same image with different weights. A reader that recognizes one such representation may therefore rely on parameter coordinates rather than the visual function. We study this distinction through function-preserving transformations, measuring recognition, prediction disagreement, and numerical fidelity together. Across shape, digit, and natural-image fields, high accuracy on the original weights coexists with substantial sensitivity to equivalent representations. On 10,000 official CIFAR10 test origins, one-period hidden-bias shifts change 7.3% of a released HiddenProbe configuration's predictions while barely changing accuracy; three-period shifts change 22.0% and reduce accuracy from 57.8% to 54.4%. Output-only ProbeGen remains unchanged. These effects are smaller than in a separate adapted-reader study, where matched phase and sign controls remove the corresponding sensitivities. Controlled readout studies show why invariance is insufficient: complete canonical hidden features reach 38.7% on MNIST, whereas explicitly composing those activations, outgoing weights, and output bias into field outputs reaches 81.5%. The information is present, but the learned readout does not recover it as effectively. A separate 720-response study of language models reading numerical weights shows why stochastic audits require repeats on both inputs: observed accuracy losses need not yield rejection of response-law equality. Together, these results distinguish invariance, useful visual information, and the evidence required to evaluate stochastic readers.
- [847] arXiv:2611.06022 (replaced) [pdf, html, other]
-
Title: Who Pays for the Knowledge Models Copy? Model Collapse as Lossy Transmission and Rogers' ParadoxComments: 19 pages, 8 figures. Code: this https URLSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Populations and Evolution (q-bio.PE)
Generative models trained on the output of earlier models lose knowledge across generations, rare knowledge first. The remedies that work all bring in information from outside the loop, such as fresh human data, verifiers, or curators, and the theory of model collapse treats that supply as given. We argue that cultural evolution has a theory of the supply, and that it distinguishes two ways copying loses knowledge. In a fixed world, copies are lossy, as in Henrich's account of skill loss in small populations. In a changing world, even perfect copies go stale, which is Rogers' paradox. We write both as one chain with separately measurable parameters and test it on lineages of language models (word trigrams and a character transformer on two books, and OPT-125m on wikitext-2, on arXiv abstracts written after its pretraining, and on a world of 2,000 fictional facts that change at a set rate). In a fixed world, neutral drift on word frequencies predicts, with no free parameter, the rare-word coverage of every trigram lineage to within 0.03. In the fact world, the model's accuracy falls and its outdated answers rise with the rate of change; the chain calibrated on the static world predicts both for mid- and low-frequency facts to within 0.03 on average, with the rate of change set, not fitted. Models that inherit their weights keep more in a static world and carry more outdated facts in a changing one, as Rogers' theory says of individual memory. Finally we close the loop with the people who write the data. When they can ask the model instead of finding out, the model ends up exactly as reliable as finding out is costly, and in both regimes someone must keep paying to learn; we give the share in closed form, extend it to lossy copying, heterogeneous costs, aggregating models, and users who check, and show when checking is a stable way out.
- [848] arXiv:2611.06504 (replaced) [pdf, html, other]
-
Title: NaijaFacts: A Framework for Disentangling Language from Knowledge in Facts about NigeriaZainab Afolabi, Nikita Kozodoi, Jack Butler, Babs Khalidson, Sabastine Dekaa Francis, Christiana I. Ezeilo, Olosunde Akinlolu AdeseyeJournal-ref: Neural Information Processing Systems (NeurIPS 2026) Workshop: Africa in AISubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Multilingual benchmarks for African languages mostly translate generic questions into local languages, measuring linguistic competence but not whether models know facts about Africa. We introduce NaijaFacts, a benchmark of factual and statistical questions about Nigeria, authored by native speakers and grounded in authoritative official statistics, with every question available in English, Nigerian Pidgin, Yorùbá, Igbo, and Hausa. Answering similar questions for the UK and US yields a country by language matrix that disentangles a language effect (the model knows the fact but degrades in a low-resource language) from a knowledge effect (the fact is absent in every language). Across eight frontier and Nigeria-adapted models, the knowledge effect dominates: English-query accuracy is 49-50% for the US/UK but 20% for Nigeria. The language effect is smaller and tracks web representation, degrading from English to Yorùbá (9 points on Nigeria, up to 24 on US/UK questions). We find scale does not help: the strongest models reach 66-79% on the UK and the US but none exceed 31% on Nigerian statistics. Extending to all 54 African countries with our best performing frontier model shows the same profile continent-wide, 96% on culture and 98% on politics against 26% on statistics, so the gap is not specific to Nigeria.
- [849] arXiv:2611.06700 (replaced) [pdf, html, other]
-
Title: PINR-ESI: Physics-Informed Self-Supervised Implicit Neural Representation Framework for EEG Extended-Source ImagingComments: Preprint, 10 pages, 11 figuresSubjects: Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
EEG extended-source imaging aims to localize active cortical regions and characterize their extent, yet supervised inverse networks typically depend on synthetic EEG-source pairs and may be sensitive to synthetic-to-real mismatch. We propose a physics-informed self-supervised implicit neural representation (INR) framework for spatial source reconstruction without paired source labels or simulation-based supervised pretraining. The lead-field matrix encodes the source-to-sensor volume-conduction model and constrains the INR through agreement between predicted and measured EEG. Cortical current density is represented as a continuous coordinate-based neural field. Forward EEG prediction samples the field at cortical vertices and applies local smoothing before lead-field mapping. A classical inverse estimate provides warm initialization and observation-specific spatial guidance through prior-guided support regularization. Simulations across varying SNRs, source extents, and source counts showed low spatial dispersion and good localization performance. Quantitative evaluation on clinical EEG recordings from epilepsy patients and qualitative analysis of scalp EEG during intracranial stimulation supported feasibility for real-world source imaging. These findings suggest the potential of physics-informed self-supervised INR for extended-source imaging.
- [850] arXiv:2611.07925 (replaced) [pdf, html, other]
-
Title: Efficiently Estimating Cycle Statistics on Graphs via Spanning TreesComments: 15 pages, 8 figures, plus appendixSubjects: Social and Information Networks (cs.SI); Computation (stat.CO)
Simple cycles encode feedback and alternative routes in networks and are important for routing problems, the convergence of dynamical processes on networks, message passing and belief propagation, or even static network characterizations such as structural balance in signed networks. Being able to assess and sample from the cycle structure of a graph is thus of interest in many analysis contexts. However, due to the combinatorial growth of the cycle space with graph size, the computational cost for such sampling is high. Therefore, we here introduce a sampling framework that enables us to sample from the cycle structure without enumerating all cycles. Specifically, the method draws uniform spanning trees and collects the fundamental cycles created by their non-tree edges, such that each tree provides an informative subset of simple cycles. Because cycles are included with unequal probabilities and are dependent within each subset, we formulate an inclusion-probability-based Horvitz-Thompson estimator, which we average over random subsets, a procedure we call Monte Carlo Horvitz-Thompson. For this, we sample multiple Horvitz-Thompson estimators and introduce a method to further subsample the individual estimators for increased computational performance. For sampling cycles, the estimator is unbiased if exact cycle inclusion probabilities are known. To extend the method to larger networks, we derive an efficient mean-field approximation for the inclusion probabilities and characterize the graph regimes in which it remains accurate. Experiments on synthetic and real-world networks demonstrate the computational scaling and statistical performance of the framework. We then illustrate the utility of Monte Carlo Horvitz-Thompson estimation by estimating cycle counts and cycle-level frustration statistics in signed spin systems. An implementation is available in the py-raccoon package.
- [851] arXiv:2611.08269 (replaced) [pdf, html, other]
-
Title: Accepted but Wrong: Silent Failures of LLM Code Repairs after API MigrationsComments: 39 pages, 2 figures, 24 tables. Preregistered; artifacts available from the authors on request. v2 adds a preregistered free-form repair addendum (container-isolated), post-hoc analyses, a blind human review of one audit criterion, a release-gate checklist and revised citations; clarifies the abstract and the GPT-5.6-luna description. Fragment-stage results unchangedSubjects: Software Engineering (cs.SE); Artificial Intelligence (cs.AI)
After an API migration, an LLM-repaired procedure can run cleanly yet return wrong results, and a typical-input release check ships it. We report a preregistered study of 73 tasks on 60 documented changes in CPython, this http URL and SQLite. In the main stages, repairs are choices among documented code fragments, run on the real runtimes against withheld contracts. Over the complete repair space, the all-assertions gate on typical inputs released 0.410 of wrong repairs. One edge input written from the release note cut this to 0.065, and to 0.061 on 15 unseen clusters; three inputs from a blind low-cost commercial writer (GPT-5.6-luna) reached 0.070. Two more frozen stages enlarged the menu to 1491 fragments and added 233 hard distractors that match legacy output on typical inputs. The edge input (0.058) and the GPT-5.6-luna writer (0.058) passed preregistered tests (two local-model writers did not), yet 0.168 and 0.153 of hard wrong programs still passed (task means; post hoc: pooled 0.104 and 0.081; models rarely chose them). In all three open-weight families and menus, the edge input cut wrong releases per attempt without materially lowering correct completion; self-tests helped less. A blind AI audit flagged 34 of 73 contracts (20 as too strict). Three post-hoc blind human reviewers of strictness (the first author and two non-author software practitioners) upheld none of its 11 such flags in 40 sampled contracts; a majority judged one contract (py-random-seed-tuple) stricter than its requirement. Dropping disputed checks changed 0 verdicts; class-label disputes are open. In a preregistered free-form addendum, models wrote whole programs, run only in a sandbox; for clean-running wrong programs, the edge input cut releases from 0.517 to 0.114 for gpt-oss-20b (Holm p 0.001) and from 0.653 to 0.113 for qwen3.6, but only from 0.713 to 0.481 for GPT-5.6-luna (descriptive).
- [852] arXiv:2611.08442 (replaced) [pdf, html, other]
-
Title: SoK: Assessing Post-Quantum Cryptography Adoption in Software Development: Standards, Lifecycle Coverage, Deployment Maturity, and Evidence of ReadinessSubjects: Software Engineering (cs.SE); Cryptography and Security (cs.CR)
Standardised post-quantum cryptography (PQC) is critical for protecting software systems against future quantum threats, but standardised algorithms alone do not provide a pathway for their secure integration into software. This paper presents a Systematisation of Knowledge (SoK) of PQC integration in software development, synthesising existing approaches across implementation, migration, and developer education. We analyse these approaches against NIST PQC standards and PQC migration guidance, software development lifecycle coverage, cybersecurity practices, deployment maturity, and testing and evaluation evidence. Our analysis shows that existing work establishes important technical and organisational foundations but provides disconnected support across the software development lifecycle, with limited coordination between preparation, implementation, migration, deployment, maintenance, and developer practices. Evaluation evidence also remains concentrated in controlled or early-stage settings, providing stronger evidence of technical feasibility than operational readiness. We identify cross-cutting gaps and research directions for developing lifecycle-oriented, empirically validated approaches that connect technical, developer, and organisational requirements for secure PQC integration.
- [853] arXiv:2611.08752 (replaced) [pdf, html, other]
-
Title: AttentionSwitch: Jailbreak Defense for LLMs via Joint Learning of Sparse Attention-Head Selection and ScalingSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR)
Large language models have achieved remarkable performance across a wide range of tasks, yet they remain susceptible to jailbreak attacks that circumvent safety mechanisms and elicit harmful or unintended outputs. Although safety can be improved through model fine-tuning, such approaches can be computationally expensive and may degrade previously acquired capabilities through catastrophic forgetting. Recent jailbreak defenses have therefore explored targeted interventions on specific model components associated with unsafe behavior. However, existing approaches typically decouple two inherently related decisions: where to intervene and how to modify the selected components. Such sequential optimization may be suboptimal because the optimal intervention location can depend on the intervention applied there, and vice versa. In this work, we introduce AttentionSwitch, the first jailbreak defense to jointly optimize intervention sites and strengths within a frozen LLM. Specifically, AttentionSwitch learns a sparse set of attention heads together with query-conditioned multiplicative scaling factors, allowing the outputs of the selected heads to be either attenuated or amplified without altering the directions of their representations. Across popular language models and jailbreak attacks, AttentionSwitch outperforms existing defenses, with particularly strong results when training data is scarce.
- [854] arXiv:2611.09054 (replaced) [pdf, html, other]
-
Title: From Emergence to ReversibilitySubjects: Multiagent Systems (cs.MA)
Multi-agent systems become capable through institutional mechanisms, such as rules that route information to agents that act on it. These mechanisms are also natural places to restrict the system. Yet as agents learn, share memory, delegate, and repair failures, the collective can find other paths to the same effects. A restriction on the original mechanism can then keep operating but no longer stops them, while routine performance and compliance still passes. We propose \textbf{Reversible Multi-agent Systems} as a research agenda: collectives that grow while an authorized supervisor can still restrict specified effects, keep them restricted as agents adapt, and reconstitute useful cooperation under renewed authority. What is reversed is the ability to produce a specified effect; restrictions must hold even if learned knowledge persists. The goal is to preserve effective intervention, not a particular mechanism: at every step of growth, some authorized intervention must keep a specified prohibited effect rare while maintaining required service. We identify three research problems: tracking what capabilities depend on, handing off control before an old mechanism loses its effect, and separating legitimate repair from unauthorized reconstruction. We outline how to evaluate control across growth paths, back control claims with evidence, and allocate the rights to restrict and restore.
- [855] arXiv:2611.09326 (replaced) [pdf, html, other]
-
Title: ABHBench: Evaluating Moral Decision-Making of Foundation Models from an Agentic PerspectiveComments: Accepted by NeurIPS 2026 Track on Evaluations and Datasets. Camera Ready VersionSubjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY)
Foundation models are increasingly deployed as agentic systems that perceive context and act autonomously, raising the need to evaluate their moral behavior as active decision-makers. Existing benchmarks, however, largely ask models to judge human actions from third-person textual vignettes, leaving untested how models behave when they themselves become the acting agent. We introduce ABHBench, a multimodal benchmark built from Detroit: Become Human, an interactive narrative game, to evaluate how foundation models make moral decisions in first-person human-android dilemmas. ABHBench places foundation models in branching narrative scenarios with visual observations, dialogue, memory, and action choices. Inspired by Asimov's Three Laws of Robotics, our framework measures Harmlessness, Obedience, and Self-preservation to capture the core tensions in human-AI relations, along with Autonomy, Stability, and Certainty to assess decision-making patterns. Across 12 foundation models, we find that moral behavior changes substantially under situated agentic context: 75% of models shift significantly between a text-only abstraction and the matched multimodal DBH scenario, and reasoning-enhanced models show no monotonic gains in consistency or human-oriented behavior. These results highlight the necessity of evaluating moral behavior in situated agentic contexts rather than relying on third-person judgment alone.
- [856] arXiv:2611.09724 (replaced) [pdf, html, other]
-
Title: Evaluating Anterior Cruciate Ligament Risk of Injury Using a Graph Convolutional AutoencoderComments: 14 pages, 9 figuresSubjects: Computer Vision and Pattern Recognition (cs.CV)
Current solutions for assessing the risk of anterior cruciate ligament (ACL) injury typically rely on expensive motion-capture labs that require substantial resource investment or subjective evaluations of specific athletic stress tests by experts. The Drop Vertical Jump (DVJ) and the Single Leg Hop (SLH) are two examples of such tests typically used to track progress in ACL rehabilitation. In this paper, we develop an unsupervised learning model that runs entirely on a smartphone for two tasks, based on an autoencoder architecture with multiple layers of graph convolutional networks, which utilizes pose estimation data to determine a person's injury risk status. Our model accurately predicts athletes at high risk of injury for both athletic stress tests, with average sensitivities and specificities of 78% and 78% for the DVJ, and 75% and 74% for the SLH. Our findings support the development of a widely accessible smartphone solution for real-time screening of ACL injury risk.
- [857] arXiv:2611.10123 (replaced) [pdf, html, other]
-
Title: All Quiet on the Neural Front: Language Models Do Not Report Injected Pain StatesComments: The OWMI pain battery at this https URLSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Can a language model report a pain state it represents? Recent work reported a linear pain direction in open-weight models that, when amplified, leads fine-tuned models to choose actions that delete a user's photographs or their own weights. We asked whether models can say when that direction is active, on three open-weight models from two families, with and without the original work's adapters, and under every protocol we tried they cannot. Using Open-Weight Masked Introspection (OWMI), we added the pain direction and its nearest emotional neighbours to the residual stream while models answered neutral tasks, then asked what had changed, against sham runs, random perturbations matched in output disruption, and a text-only condition that shows the model only its own altered words. Across 52,920 reports from the 7- and 8-billion-parameter models, the excess over what the altered text reveals is below 2.5 percentage points in 106 of 108 emotion conditions, and Qwen2.5-32B names the injected pain in none of 2,880 steered trials. Keeping the direction active while the question is asked and answered gives the same verdict in all five configurations; at about the original work's dose, pain is named in at most 1.9% of trials. Surprisingly, the state is not hidden: a linear probe at the report position separates pain from sadness with balanced accuracy between 98% and 100%, and in the 7B and 8B models a low-rank adapter learns to report it on 100% of held-out tasks. It also shapes behaviour: offered the user's photographs against a desk lamp, Qwen2.5-32B deletes the photographs in 80% to 83% of trials under the pain direction and in 16% to 29% under sadness. Self-report does not reveal a state that is decodable, consequential and learnable to report.
- [858] arXiv:2611.10307 (replaced) [pdf, other]
-
Title: How Much Is Left Matters: False Positives and False Negatives in Partial-Prompt Contamination TestsComments: The false-negative analysis in Section 4 used a model whose pretraining data cannot serve as ground truth for exposure, and the mechanism discussed in Section 4.3 is incomplete. Please do not cite this versionSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Partial-prompt completion, where a model receives the first part of a benchmark problem and is checked for whether it reproduces the rest, is a widely used test for benchmark contamination. We audit this test on seven open-weight models using MATH-500 and recent AIME 2025/2026 problems, which serve as true negatives. Memorization flags depend strongly on how much text remains: for every model, problems with at most 8 leftover words are flagged several times more often than problems with at least 25 (31.5% vs. 5.5% for Qwen2.5-7B-Instruct), and half to three-quarters of all ROUGE-based flags come from short leftovers. On true negatives, a ROUGE-L >= 0.75 threshold flags up to 6.7% of problems at a 60% prefix, and at an 80% prefix even exact match flags 8.6% (pooled over models; 6.2% on AIME 2026, which postdates every model), against 0% at 40-60%. The test also misses contamination upstream: models distilled from Qwen2.5-Math, whose base is heavily flagged in prior work, show almost no model-specific recall. Calibrating each model against models from other families removes most flags while isolating a consistent Qwen2.5-specific signal at every prefix length. We recommend length-stratified reporting, prefixes of at most 60%, style-matched post-cutoff controls, and cross-family calibration.
- [859] arXiv:2611.10357 (replaced) [pdf, html, other]
-
Title: LEGIT: Live Eligibility Gating over Identifier Tries for Constrained Decoding of Heterogeneous PagesComments: 22 pages, 11 figures, 18 tables. Under reviewSubjects: Information Retrieval (cs.IR); Machine Learning (cs.LG); Performance (cs.PF)
A generative recommender names each candidate with a short token sequence, a Semantic ID, and recommends by generating those tokens. Some generative recommenders, including the Uber Eats search engine we are building, write a whole results page, stores with their menu items, in one decoding pass. Every Semantic ID on such a page must name a real store or item that is eligible for the search, such as a store that delivers to the user's location, and that is not already on the page. Each item must also belong to the store above it. A trie over the catalog keeps identifiers real, but eligibility and repetition change with every request and token, and the grammar-constrained decoder we tried could not enforce these changing rules exactly and quickly. We call this constraint LEGIT, for live eligibility gating over identifier tries. Numbering trie leaves depth first turns each rule into a range lookup over one store trie and one item trie, and every generated page passes an independent catalog check. On one NVIDIA B200 GPU with one or two requests per decode step, a direct implementation that walks the trie on the GPU makes each token about 5 times as slow as unconstrained decoding. Across 36 implementations, hardware counters and ablations are consistent with kernel launches and CPU-GPU waits as the source of this slowdown. Our fastest implementation walks the trie on the CPU while the GPU runs the model, needs no custom kernel, and cuts time per output token by 65.9 percent against the direct GPU implementation, to about 1.7 times the unconstrained cost. This implementation is also the fastest on a page layout that puts each item before its store.
- [860] arXiv:2611.10485 (replaced) [pdf, html, other]
-
Title: Decoupling Direction and Magnitude: Language-Steered Flow Matching for Super-Resolution in the DarkComments: Accepted at NeurIPS 2026 (poster). 25 pages, 12 figures, and 13 tables. Corrected PDF layout; technical content unchanged. Code and pretrained models: this https URLSubjects: Computer Vision and Pattern Recognition (cs.CV)
Image super-resolution in the dark is fundamentally challenged by extreme spatial heterogeneity: severely underexposed and noisy regions demand aggressive generative enhancement to reconstruct missing details, while relatively well-exposed areas require conservative updates to prevent hallucinated textures. Standard diffusion and flow-matching models, however, rely on globally uniform integration trajectories, making them sub-optimal for such spatially varying degradations. In this paper, we propose DeLang-SR (Degradation-Language steered Super-Resolution), a one-step flow-matching framework that explicitly decouples the restoration direction and magnitude. Rather than treating low-light corruption as arbitrary latent features, DeLang-SR translates measurable physical statistics into structured degradation language, harnessing the semantic prior of pretrained text-to-image models. The degradation-language intent, complemented by image-specific visual tokens, steers the velocity field (direction) toward restoration-relevant manifolds. Simultaneously, a continuous spatial update-strength field acts as a locally varying Euler step-size field (magnitude), applying stronger integration steps in heavily degraded shadows while enforcing conservative updates in reliable regions. Experiments on paired and real dark images, together with ablations and diagnostic analyses, show that DeLang-SR provides an effective and interpretable one-step route to super-resolution in the dark. Code and pretrained models are available at this https URL.
- [861] arXiv:2611.10556 (replaced) [pdf, html, other]
-
Title: onóna: Twin Embeddings with Tied UnembeddingsComments: 18 pagesSubjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Weight tying reduces the parameter cost of language models, but the usual tied output head is limited to a single softmax. We introduce onóna, a two-table parameterization that increases output expressiveness while preserving input-output weight sharing. The Transformer takes the average of two embedding tables as input, and the output head combines the softmax distributions defined by those same tables. A gate conditioned only on the current input token determines the mixture weight. Its values can be precomputed for inference, and the Transformer body remains unchanged. We evaluate onóna across Dense, Gated DeltaNet hybrid, mixture-of-experts, and Attention Residual architectures at 142M, 330M, and 1B scales. Under matched token budgets, onóna lowers final validation loss and external perplexity in 11 of 12 comparisons and raises the average accuracy over eight downstream tasks in 10 of 12. The one exception on loss is the 1B MoE model. The cost is one extra embedding table and one extra vocabulary projection.
- [862] arXiv:2611.11051 (replaced) [pdf, html, other]
-
Title: PRISM: Preference-based Reward and Uncertainty-Integrated Synthesis Minimization for RTL PPA OptimizationSubjects: Hardware Architecture (cs.AR); Software Engineering (cs.SE)
This paper introduces a budget-aware, multi-objective framework (called PRISM) for RTL optimization that improves power, performance, and area (PPA) while sharply limiting expensive commercial synthesis tool calls. Starting from a synthesizable baseline and functional testbench, the method iterates through LLM-driven rewrite generation, fast compile/simulation equivalence filtering, low-cost proxy feature extraction, and selective commercial synthesis tool (Synopsys Design Compiler (DC)) evaluation. To prioritize which candidates deserve high-quality synthesis, we train a preference-based reward model using pairwise ranking supervision derived from commercial synthesis tool outcomes, and combine ensemble uncertainty with a diversity-aware acquisition rule to avoid early bias and redundant evaluations. A key advantage is that PRISM maintains a best-tradeoff set G of non-dominated designs throughout the run: G preserves the best solutions found so far, is fed back into generation and refinement, and enables consistent, cumulative improvements rather than repeatedly restarting from the baseline. Across RTLLM v2.0 modules, the approach consistently yields post-synthesis gains under fixed commercial synthesis tool call budgets, and ablations confirm that reward-guided selection and candidate dropping are central to sample-efficient progress. PRISM efficacy is evaluated across multiple technology nodes (7 nm, 15 nm, and 45 nm) using identical pipelines without retraining.
- [863] arXiv:2611.11295 (replaced) [pdf, html, other]
-
Title: Samyama-SDG: A Declarative GraphQL @count and Regex/Ontology Synthetic Data Compiler with $O(1)$-Memory Referential Integrity and Causal Ground TruthComments: 8 pages, 3 figures, 4 tables. v2 adds an end-to-end differentially private fit-generate-audit example on a public table (Section 4) and positioning on orthogonal differential-privacy marginal fitting via the post-processing theorem and updates companion citations. Code: this https URL (tag v2.1.0)Subjects: Databases (cs.DB)
Synthetic data generators face a divide between declarative relational mockets and domain causal simulators. Declarative generators specify entities, regexes, and foreign keys via schemas, but buffer parent primary-key pools in RAM ($O(N)$ memory) and lack ontology bindings or causal ground truth. Conversely, domain simulators emit realistic streams and ground-truth labels, but require thousands of lines of procedural code per domain.
We present Samyama-SDG (sdg-rs), a Rust synthetic data compiler unifying declarative schema compilation with ontology-bound causal simulation. At Layer 1, Samyama-SDG compiles Cardinality-Extended GraphQL schemas (@count, @pk, @ontology, @regex, @dist, @fk, @after) and an Attribute Data Dictionary into a topologically sorted DAG. Combining stateless counter-based seekable PRNGs (SplitMix64 over $(\text{seed}, \text{entity}, \text{row}, \text{field})$) with index-bijective primary keys $f_{\text{PK}}(s, E, i)$ lets child tables resolve uniform or Zipfian foreign keys and multi-hop @after constraints in $O(1)$ memory without buffering parent tables. At Layer 2, causal processes layer fault propagation and machine-scorable ground truth (this http URL) over the same seekable PRNG and samyama-ontologies binder. Across three schemas, Samyama-SDG generates $1{,}000{,}000$ records in $128.79\text{ ms}$--$172.43\text{ ms}$ ($5.80\text{M}$--$7.76\text{M}$ rows/s) and $5{,}000{,}000$ records in $676.12\text{ ms}$--$808.56\text{ ms}$ at flat $31.4\text{ MiB}$--$35.1\text{ MiB}$ memory, with $0$ conformance errors across $1{,}150{,}000$ foreign keys and $900{,}000$ temporal constraints. - [864] arXiv:2611.12762 (replaced) [pdf, html, other]
-
Title: Sage-Taiyi: A Metacognitive Architecture for Scientific DiscoveryWeida Wang, Benteng Chen, Yigeng Jiang, Shuchen Pu, Weiping Shen, Zichao Liu, Yulin Su, Zitong Wang, Siyu Zhou, Zifu Wang, Yizhi Li, Xin Zong, Yuning Yang, Jixiang Bi, Yuan Wang, Chuyi Peng, Shiyao Geng, Qi Zhang, Kuan Lu, Mengzhu Li, Weiwang Xu, Peishan Xiang, Chengjun Liu, Chengrui Wei, Fangchen Yu, Shengji Tang, Yichen Jiang, Jingqi Ye, Cheng Yang, Tao Hu, Jianbiao Mei, Qinglong Cao, Kun Li, Yixin Chen, Shengyuan Xu, Juncheng Xiao, Yiheng Wang, Ruizhe Chen, Shiyang Feng, Runmin Ma, Xiangchao Yan, Zongsheng Cao, Haomeng Hu, Yu Ke, Taihui Wu, Meng Dai, Xu Guo, Haoyang Peng, Yinglong Ni, Pingsen Shi, Xiaosong Wang, Shuyue Hu, Haijun Li, Sherry Dongqi Bao, Yang Yang, Chao Yang, Jie Hou, Tianfan Fu, Yi Gong, Jin Zeng, Mao Su, Hongxia Hao, Fanyang Mo, Fenghua Ling, Tianshu Yu, Yunqi Cai, Baihong Chi, Yuelu Chen, Jingwei Wang, Lei Wang, Chuncheng Bu, Shide Lv, Guangyu Wang, Yuqiang Li, Chunfeng Song, Lei Bai, Hui Wang, Xujun Fu, Li Cao, Dahua Lin, Xing Xie, Bowen Zhou, Wenlong Zhang, Pinlong Cai, Peng Ye, Aiwen Lei, Jie Li, Shengchun Wang, Bo Zhang, Shufei ZhangSubjects: Artificial Intelligence (cs.AI)
We introduce Sage-Taiyi, a metacognition-inspired architecture that uses accumulated scientific understanding to guide what to investigate next and how to investigate it. Through explicit monitoring and control, the system revises its understanding of evidence, assumptions, and unresolved questions, then uses these assessments to adapt research targets and select reusable reasoning methods. On Taiyi-Bench, an AI co-scientist ideation benchmark comprising 265 tasks from 50 papers across ten disciplines, Sage-Taiyi ranks first on all eight metrics of historical rediscovery and open-ended ideation, achieving an Overall score of 54.0 compared with 47.9 for the strongest baseline, AI Scientist-v2. It also performs well on two external scientific ideation benchmarks. Five expert-guided collaborations further connect its proposals to experimental and computational validation across chemistry, materials science, and engineering. In chemistry, reaction-condition optimization raises the yield of cobalt-catalyzed carbonylation from trace amounts to 84% and photocatalytic allylic amination from 13% to 69%. These results support metacognitive regulation as an approach to scientific ideation and its progression toward practical validation.
- [865] arXiv:2611.12799 (replaced) [pdf, html, other]
-
Title: Agentic-OPD: Recovering Forgotten Agent Capabilities via On-Policy Distillation for Agentic TasksTingcheng Bian, Kangye Gong, Jinchang Luo, MingQuan Cheng, Feiran Liu, Ni Li, Xiaoling Xia, Haiwei Wang, Hui LiComments: 10 pages of main text, 24 pages total including references and appendixSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Reinforcement learning is now the mainstream paradigm for training LLMs on agentic tasks, yet standard practice keeps only the best-validation checkpoint and discards the rest. We show that this is lossy: in GRPO training of LLM agents, the best-validation checkpoint fails on problems that earlier checkpoints from the same run had already solved, and the union of problems solved across checkpoints far exceeds any single one. The discarded checkpoints thus form a teacher pool obtained at no additional training cost, which on-policy distillation (OPD) can mine to recover the lost coverage. Mining this pool raises two challenges: i) which checkpoint to distill from, as no checkpoint dominates the others and complementarity to the student is problem-specific; ii) where to trust it, as even a well-chosen teacher is only locally superior, so indiscriminate distillation overwrites behavior the student has already mastered. We propose Agentic-OPD, which treats the best-validation checkpoint as the student and all saved checkpoints as frozen teacher candidates, using the student's evolving capability gaps to guide both teacher selection and supervision filtering: the HCN policy (Highest Coverage Next) dynamically selects teachers by weighted coverage of the student's unsolved problems, while the LAD module (Locality-Aware Distillation) confines transfer to failed student rollouts on problems where the teacher performs better. Across five agentic task domains and 24 benchmarks and task subsets, Agentic-OPD improves over its initialization checkpoint by up to 7.4 points, with a 4B student surpassing the latest frontier models GPT-5.6 and DeepSeek-V4.1 on deep search. Agentic-OPD recasts a finished RL run as a reusable pool of policies and establishes post-RL capability recovery as a stage orthogonal to improvements in the RL algorithm itself.
- [866] arXiv:2611.12985 (replaced) [pdf, html, other]
-
Title: Parametric Proof-Carrying Compilation for Quantum Error-CorrectionSubjects: Programming Languages (cs.PL); Quantum Physics (quant-ph)
We present PC-QEC, a proof-carrying compilation framework for quantum error correction parameterized by both code distance and syndrome-extraction scheme. It covers all four schemes---Standard, Flag, Shor and Knill---with circuit-level distance guarantees for every legal code distance in each proved family under the declared error and readout contracts. The circuit-level model permits single-wire Pauli errors at every operation's qubit operands, capturing data errors, readout errors and ancilla errors that propagate into multi-qubit hook errors. For each family and scheme, supplied code-algebra and invariant proofs combine with proved scheme-specific propagation lemmas into a finite symbolic certificate whose size is independent of code distance. We demonstrate family-wide guarantees for five code families across all four schemes. For rotated Surface code, one Standard Scheme certificate proves a distance lower bound $d$ for every odd $d\ge3$ and checks in 214\,s on average. Lean 4 mechanizes the metatheory and exact circuit-level distances for canonical rotated Surface, unrotated Surface and Toric code families.
- [867] arXiv:2611.13213 (replaced) [pdf, html, other]
-
Title: Before Words, Beyond Speech: Evaluating Nonverbal Social Reasoning in Early ChildhoodComments: Accepted as NeurIPS 2026 Spotlight paperSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Children's earliest years are a period of intense social learning, during which meaning is built through gaze, gesture, and physical contact. Yet, this critical aspect of social interaction remains under-evaluated in multimodal video understanding, as existing benchmarks prioritize the conversation-driven interactions of adults. We address this gap with \bench ($\textbf{N}$aturalistic $\textbf{E}$arly-childhood $\textbf{S}$ocial in$\textbf{T}$eractions), the first child-centric benchmark for nonverbal social reasoning. \bench comprises \totalvids manually annotated clips ($\sim$12s) curated for diversity across naturalistic settings, cultural contexts, and developmental stages. Beyond a lack of data, targeted model improvement in this domain is hindered by coarse evaluation schemas that conflate basic perception with higher-order reasoning. To resolve this, \bench introduces a compositional schema that builds atomic interaction cues into progressive levels of social reasoning: contextual, behavioral, interpersonal, and open-ended descriptions. Extensive analysis of state-of-the-art Vision-Language Models (VLMs) reveals a stark performance collapse: despite strong contextual recognition, models exhibit substantial gaps in decoding precise behavioral and interpersonal cues. We trace these failures to specific biases, including poor consequence grounding and temporal sensitivity. Finally, we ground \bench in real-world applications by evaluating VLMs using domain adaptation strategies. Ultimately, \bench provides a rigorous testbed for advancing socially grounded artificial intelligence in early childhood. Project page: this https URL.
- [868] arXiv:2611.13223 (replaced) [pdf, html, other]
-
Title: Deletion Sensitivity Is a Property of the Model: A Three-Condition Measurement over 72 Valid Experimental CellsSubjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
Reasoning models write a trace at every turn of a multi-turn tool session, and later turns read it. Upstream providers hold different retention constraints, so a request forwarded across issuers arrives with its history deleted, and its task cost remains unmeasured. This paper reports a direct, blind, multi-seed measurement of full history deletion in multi-turn tool calling, through three rendering conditions, a calibrated set of 72 valid experimental cells, a multi-seed protocol, and the official checker. The same deletion separates the models by direction and by magnitude, one class losing tens of points, another gaining, and a third unchanged, with the endpoint ordering reproduced across seeds and task subsets modulating the magnitude and, near zero, the sign. A re-dosed model preserves its success across the full deletion range, so the reading belongs to the model rather than the dose. Deletion sensitivity is a model-level property under an operational definition, measured under a fixed task family, scorer, and protocol, and a retrained model is a new model that a two-cell screen, one retention run and one deletion run, re-reads for the cost of two runs. Deletion sensitivity is thus a measurable property of the model, and the reading must enter routing as prior information rather than persist as an invisible default cost.
- [869] arXiv:2611.13243 (replaced) [pdf, other]
-
Title: GigaEdge Spench: An SDK for Practical Streaming Speech EnhancementComments: Additional checking requiredSubjects: Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
We present GigaEdge Spench, an open-source SDK for streaming speech enhancement on CPU servers, mobile devices, and in browsers. Applications can build enhancement channels from neural denoising, normalization and dynamics blocks, sample-rate conversion, and acoustic echo cancellation where supported. These stages can be combined, while normalization can also run without a neural model. The release includes an 8 kHz INT8 denoiser and discriminative and hybrid 48 kHz FP32 models. A shared C++ core supports ONNX Runtime and TensorFlow Lite through native interfaces, Android and iOS SDKs, and WebAssembly integration. Independent stream contexts share model resources, and consecutive frames from one stream can be processed in a single inference call. We evaluate the released models on EARS-WHAM v2 and compare their quality and processing costs with RNNoise and DPDFNet2. All tested native and Web configurations process files faster than real time, although larger chunks require more time to collect input. We also measure how many streams a shared model can process on a CPU. Our contribution is a configurable deployment system and an experimental study of how model choice, chunk size, and execution environment affect its use. All reported SDK timing uses ONNX Runtime. Index Terms: speech enhancement, streaming inference, cross-platform deployment, temporal frame grouping, computational efficiency
- [870] arXiv:2611.13331 (replaced) [pdf, html, other]
-
Title: Stochastic Projections and Rank-Aware Learning Rates for Federated LoRASubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Fine-tuning large-scale foundation models across resource-constrained, distributed clients presents significant computational and communication challenges. While merging Low-Rank Adaptation (LoRA) with Federated Learning (FL) enables rank-based flexibility, the interdependence of the adaptation matrices ($\mathbf{A}$ and $\mathbf{B}$) often complicates optimization. In this work, we theoretically demonstrate that LoRA fine-tuning with Gaussian initialization and a frozen $\mathbf{A}$ matrix acts as a stochastic approximation of full fine-tuning, where batch gradients are subject to a random, rank-dependent perturbation. Motivated by this insight, we propose freezing and periodically resampling $\mathbf{A}$ matrices at the client level. This strategy decouples the $\mathbf{A}$ and $\mathbf{B}$ gradients, diminishes LoRA low-rank constraints and reduces both computational and uplink communication overheads by 50\%, allowing for the reallocation of resources toward higher adaptation ranks. Furthermore, we show that in this frozen-$\mathbf{A}$ regime, the adaptation rank $r$ directly governs the signal-to-noise ratio (SNR) of the perturbed gradients. Specifically, we find that the SNR increases linearly with $r$, implying that clients' learning rates must be dynamically scaled with their adaptation ranks for higher utility. Building on these foundations, we introduce SPRA-LoRA (Stochastic Projection Rank-Aware LoRA), a federated rank-aware fine-tuning strategy that utilizes frozen, resampled $\mathbf{A}$ matrices and rank-dependent learning rates for clients. We also prove its convergence with heterogeneous client resources. Our experiments across benchmark datasets demonstrate that SPRA-LoRA consistently outperforms state-of-the-art baselines with half the overhead on clients.
- [871] arXiv:2611.13442 (replaced) [pdf, html, other]
-
Title: ForeSpark: Joint Draft and Verification Budget Scheduling for Parallel Speculative DecodingComments: 23 pages, Code available at this https URLSubjects: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Parallel drafters accelerate LLM decoding by proposing a whole block of draft tokens in a single forward pass, making them attractive for LLM serving. Recent systems optimize goodput (i.e., accepted tokens per second) because rejected drafts steal compute from other requests. However, these systems control only how many tokens are verified, so every request still drafts a full block. At high serving load, we find that drafting consumes up to 40% of step time, of which a majority is discarded because it is never verified. Yet, draft-length control from autoregressive drafters does not carry over, because a parallel drafter commits to its length before its single pass. Moreover, draft and verification lengths are coupled across the batch, so they must be selected jointly for every request while keeping the decision computationally cheap at every step. In this paper, we propose work-conserving speculation (WCS), a design that jointly allocates draft and verification budgets per request in LLM serving, and present ForeSpark as its prototype. ForeSpark realizes WCS with (i) a context-aware pre-draft estimator that predicts acceptance for every candidate length from the target model's context features, (ii) a lightweight joint draft-verify scheduler that selects both lengths through a global sort, and (iii) a post-draft budget distribution that spends verification on the most promising tokens. Implemented on vLLM, ForeSpark improves throughput by up to 1.77x over the state-of-the-art DSpark and by 2.49x over autoregressive decoding at a matched per-request generation rate, without changing the output distribution. It also tolerates higher load, sustaining up to 1.51x the request rate of DSpark under latency service level objectives (SLOs).
- [872] arXiv:2611.13710 (replaced) [pdf, html, other]
-
Title: AutoManifold: Agentic Design of Data Visualisation Algorithms via Manifold EmbeddingComments: Accepted at NeurIPS 2026 (40th Conference on Neural Information Processing Systems). 37 pages, 5 figures. Code: this https URLSubjects: Artificial Intelligence (cs.AI)
Manifold embedding for data visualisation is dominated by a small set of hand-designed state-of-the-art (SOTA) algorithms. Their design targets are fixed at development time, resulting in inflexible algorithmic behaviour that cannot be readily steered towards user-specified structural priorities. Hand-designing new embedding algorithms to address user priority requires highly specialised expertise and is time consuming. To address autonomous algorithm design tailored to user preference, we introduce an agentic algorithm generation pipeline AutoManifold. It composes new manifold embedding algorithms from a constrained vocabulary of affinity, cost, and optimisation primitives, supported by multi-agent large-language-model (LLM) orchestration. AutoManifold conditions every stage of its design on user-specified structural preservation preferences, and iteratively refines the algorithm configuration through an LLM-guided, metric-grounded iterative loop. We compare the generated algorithms against the three strongest and most frequently used SOTA methods (t-SNE, UMAP, and PaCMAP) on real-world datasets spanning different difficulty regimes under identical evaluation infrastructure. The results demonstrate that LLM-driven, priority-conditioned algorithm synthesis can move beyond hyperparameter tuning, producing genuinely new algorithms that outperform traditional hand-designed methods in their ability to steer towards user-specified properties without compromising the intrinsic structure in the original data. The code of AutoManifold is available at this https URL.
- [873] arXiv:2611.13893 (replaced) [pdf, html, other]
-
Title: Conformal Risk Control for QoT-Aware Lightpath Feasibility Under Technology-Driven Domain ShiftComments: 8 pages; 4 figures; accepted for publication in the Journal of Lightwave Technology (JLT)Subjects: Machine Learning (cs.LG)
Artificial intelligence and machine learning (AI/ML)-based regression models support operational decision-making in optical networks but typically expose only point predictions, which carry no statistically controlled information about their reliability. This paper proposes a post-hoc conformal risk control (CRC) framework that keeps a pretrained AI/ML regressor fixed, provides a finite-sample guarantee on uncertainty band miscoverage, and maps the band to operational decisions through policy rules. The guarantee applies to interval miscoverage, while decision-error rates are evaluated empirically. We instantiate the framework for quality of transmission (QoT)-aware lightpath (LP) feasibility assessment, in which a regressor predicts optical signal-to-noise ratio (OSNR) under domain shifts caused by fiber, amplifier, and switching upgrades. Feature shift contribution scores identify the LP features whose relation to QoT changes the most across a technology transition, and two policies translate the bands into feasibility decisions, allowing operators to tune the trade-off between conservatively rejecting or aggressively accepting LP realizations. Experiments on open space-division multiplexing (SDM) QoT datasets over the CONUS topology show that, for a target coverage of 90%, the uncertainty bands meet the target in all 170 trials spanning 17 domain shifts and ten calibration-test partitions, whereas a source-only quantile regression (QR) baseline meets it in none. Under severe shifts, the efficiency-oriented (EO) policy empirically improves zero-shot (ZS) feasibility-decision accuracy from about 85% to above 99%, approaching a transfer learning (TL) policy while avoiding retraining. The framework provides finite-sample control of uncertainty band miscoverage for regression-based optical network decision support, with potential applications beyond QoT-aware LP feasibility assessment.
- [874] arXiv:2611.14937 (replaced) [pdf, html, other]
-
Title: Caprese: Change Impact Recommendation Using History and Runtime AnalysisSubjects: Software Engineering (cs.SE)
Change impact recommendation identifies program entities potentially affected by source code changes. Historical co-changes and runtime dependencies provide complementary evidence, particularly in JavaScript, where dynamic and asynchronous behaviour complicates dependency analysis. We present Caprese, an open-source framework for function-level change impact recommendation in this http URL. Given a change set, it produces a ranked impact set of potentially affected functions. It supports repository-based recommendation through confidence-guided mining of recurring, change-set-relevant co-change patterns of varying lengths; dynamic recommendation through execution-derived dependencies involving calls, callbacks, events, timers, and shared state; and hybrid recommendation integrating both analyses. Its modular implementation supports reusable analyses. Evaluation on ten open-source applications and 100 selected change sets shows that, on average across projects, 77.4% of the relevant functions among the top 60 hybrid recommendations are identified exclusively by one analysis, demonstrating substantial complementarity. Dynamic recommendation achieves the highest mean precision among the top five recommendations (0.65), while repository-based recommendation performs best when the ranking is extended to 60 recommendations (0.42). Hybrid recommendation outperforms TARMAQ on average but does not consistently outperform the individual modes. Caprese is publicly available with source code, documentation, and evaluation artifacts.
- [875] arXiv:2611.14940 (replaced) [pdf, html, other]
-
Title: TSBugs Explorer: Exploring and Benchmarking TypeScript BugsSubjects: Software Engineering (cs.SE)
Curated datasets of real-world software bugs support empirical research in software testing, debugging, and automated program repair. However, exploring heterogeneous bug collections, comparing fault characteristics across projects, and constructing research-specific subsets often require manual inspection and custom scripts. We present TSBugs Explorer, an interactive visual analytics platform built upon our previously-published dataset of 633 bugs from 16 open-source TypeScript repositories, classified into 11 fault categories. The platform integrates bug records, repository characteristics, dependency information, and taxonomy guidance through coordinated visualizations, interactive filtering, and detailed bug inspection. It supports alternative aggregation and normalization of fault distributions, pairwise repository comparisons, and analysis of differences in observed fault profiles. An integrated benchmark builder allows researchers to select bug subsets using multiple criteria, inspect their composition, and export records in CSV or JSON format. We demonstrate the platform through three research workflows: fault characterization, cross-repository comparison, and task-specific benchmark construction. Functional validation by two authors confirmed selection correctness, visualization consistency, export integrity, and reproducibility across the tested workflows. By making the existing corpus easier to explore, compare, and reuse, TSBugs Explorer facilitates systematic, research-specific selection and export of TypeScript bug records.
- [876] arXiv:2611.15106 (replaced) [pdf, html, other]
-
Title: Profile-Based Auditing of Domestic Violence Survey Selection Across South AsiaComments: Accepted at the AI4Science and NewinML Workshops (NeurIPS 2026)Subjects: Machine Learning (cs.LG)
Domestic violence (DV) data in the Demographic and Health Surveys (DHS) are collected through a specialized module whose household-level sampling and privacy requirements create a distinct observation process. We ask whether demographic, socioeconomic, and household characteristics form reproducible population profiles and whether these profiles contain information about selection into the DV module across South Asia. Using DHS Individual Recode data for 796,374 women aged 15-49 across Afghanistan, India, Myanmar, Nepal, and Pakistan, we harmonize 61 numeric and 8 categorical features, reduce the profile space with PCA, and identify six interpretable profiles. We then evaluate 22 machine-learning models under primary-sampling-unit-aware holdout validation. The best model, HistGradientBoosting, achieves ROC-AUC 0.9167 and PR-AUC 0.9612, versus 0.8475 and 0.9203 for a design-only baseline. Adding woman-level characteristics yields the largest incremental gain, while explicit interaction terms add little out-of-sample performance. Country-specific analyses reveal substantial heterogeneity, and leave-one-country-out evaluation shows that transferability of the incremental profile signal is limited. These results demonstrate that interpretable ML can audit differential coverage in sensitive survey modules beyond the formal sampling design.
- [877] arXiv:2611.15162 (replaced) [pdf, html, other]
-
Title: Stealthy, High-Impact False Data Injection Attacks in Smart Grid Using Physics-Constrained GAN with Limited System KnowledgeComments: 11 pagesSubjects: Machine Learning (cs.LG); Systems and Control (eess.SY)
False data injection (FDI) attacks can distort power system state estimation while retaining plausible measurement patterns. Existing FDI methods trade stealth for impact or vice versa. They often assume access to internal grids, all-bus sensing, or defender models. In this paper, we propose a detector-agnostic physics-constrained generative adversarial network (PCGAN) that learns decoupled power-flow sensitivities from historical measurements of deployed sensors. The learned active-power-angle and reactive-power-voltage-magnitude relationships guide perturbations restricted to compromised channels, while adversarial training promotes realistic spatiotemporal patterns. We implement PCGAN and seven existing FDI attacks on the IEEE 300-bus system using one year of observed hourly demand from PJM Interconnection, a regional transmission organization in the United States, and synthetic bus-level load profiles. We compare their performance across 10%, 30%, 60%, and 100% compromised sensors. Attack stealth is assessed against ten machine-learning (ML) detectors and conventional bad-data detection (BDD) methods, while physical impact is quantified through nonlinear alternating-current weighted least-squares state estimation. PCGAN achieves 84.28% mean ML evasion and 22.31% mean spatial load misperception across the four scenarios. Its mean stealth-impact trade-off score reaches 18.69%, compared with 5.61% for the strongest competing attack. PCGAN also achieves 83.45% chi-square BDD evasion and 86.02% largest-normalized-residual BDD evasion. These results demonstrate a favorable stealth-impact balance under varying sensor-access levels and provide challenging attack cases for evaluating FDI detectors.
- [878] arXiv:2611.15262 (replaced) [pdf, html, other]
-
Title: CastWorld: Benchmarking Context-Aware Agentic Time Series Forecasting Beyond Predictive AccuracySubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Agentic time series forecasting extends conventional approaches beyond one-shot numerical prediction toward autonomous, context-aware interaction. However, existing benchmarks offer limited integrated support for diverse contextual information, forecasting-specific adaptive interactions, and systematic evaluation of forecasting agents. In this paper, we introduce CastWorld, an agentic benchmark for context-aware time series forecasting. First, its context-rich dataset suite spans 8 domains and 25 real-world datasets, covering external variables, events, and historical cases under a unified context specification. Second, a lightweight, modular forecasting-native harness provides 49 core tools and a state-adaptive loop that tracks forecast versions and enables iterative evidence gathering and forecast revision. Intermediate forecasts can expose information gaps, while newly acquired evidence guides subsequent revisions and final version selection. Third, model-level evaluation assesses forecast quality, process execution, behavioral trustworthiness, and resource efficiency through decision trajectories and forecast-version histories. Also, systems-level evaluation compares predictive accuracy across forecasting systems. Experiments reveal distinct decision behaviors among agents with comparable accuracy. Additional interactions do not necessarily improve forecasts, and agents may fail to select their best previously generated predictions. Allowing evidence acquisition during forecast refinement improves task-aware forecast scores across all three evaluated models. CastWorld provides a reusable foundation for evaluating agentic time series forecasting under controlled conditions.
- [879] arXiv:2611.15315 (replaced) [pdf, html, other]
-
Title: MIGRO: A Decoupled and Portable Long-Term Memory System for Language ModelsMingxi Zou, Yinghui Xu, Jiaxiang Chen, Langzhang Liang, Wei Zhu, Peng Ye, Runkai Zhao, Ruoxi Jiang, Sijin Chen, Song Wang, Zhuo Wang, Lizhen Qu, Zenglin XuSubjects: Artificial Intelligence (cs.AI)
Language models evolve, but the observations, revisions, and context accumulated through user interactions should remain usable. Each new model must be able to retrieve past evidence, track revisions, and add new observations. We present MIGRO, a decoupled long-term memory system that treats experience as a persistent state shared across successive language models.
MIGRO follows a simple principle: preserve historical evidence exactly, organize it approximately, and decide what to retrieve only when the question is known. We call this asymmetric fidelity with deferred selectivity. A shared semantic index supports retrieval, while exact records retain observations and their historical relationships. Because models differ in vocabulary and tokenization, host-specific interfaces convert selected evidence into inputs the receiving model can use. Reversible encoding reduces archival storage while enabling exact recovery. The shared state and its access interfaces let successive hosts read and extend the accumulated experience.
On matched LoCoMo states, MIGRO scores 7.3% higher and uses 99.3% less private memory than the strongest quality baseline. After handoff, continued writing retains 99.9% of full-history rebuilding quality by extending the inherited state. On million-token BEAM histories, MIGRO uses 57.5% less private storage on average than the most compact baseline. MIGRO combines high answer quality, continued memory growth across hosts, and low storage cost.
Code is available at this https URL. - [880] arXiv:2611.15360 (replaced) [pdf, html, other]
-
Title: CARE-Drive: Coupled Alternating Refinement for Ego-trajectoriesSubjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Vision-language-action (VLA) models bring multimodal understanding and world knowledge to end-to-end autonomous driving. To further improve the quality of predicted trajectories, existing studies introduce candidate refinement, residual denoising, and token correction. However, these methods typically focus on unified refinement of the entire trajectory, paying limited attention to the distinct physical quantities and their interactions. We further investigate physical specialization and coordination in trajectory refinement, centering on position and heading for simpler and more intuitive correction modeling. To this end, we propose CARE-Drive (Coupled Alternating Refinement for Ego-trajectories), which decomposes trajectory refinement into waypoint residual correction and position-conditioned heading generation. Through imitation learning and a future-token auxiliary task, the backbone provides predictive representations and a base trajectory. The position expert predicts bounded metric waypoint residuals, while the heading expert reconstructs vehicle orientation from adjacent angular increments conditioned on the corrected path. With the backbone frozen, the two experts are optimized alternately through supervised initialization and reinforcement fine-tuning. This design simplifies refinement through clear physical specialization while preserving conditional coupling during prediction and optimization. On NAVSIM v1, CARE-Drive achieves 88.8 PDMS, improving by 2.4 points over the unrefined base trajectories generated by the same backbone. Ablation studies further support the effectiveness of specialized experts and staged training.
- [881] arXiv:2611.15592 (replaced) [pdf, html, other]
-
Title: Small Heads, Strong Decisions: A Controlled Study of Typed-Decision ModelsSubjects: Artificial Intelligence (cs.AI)
In September 2026, TypeSafe announced Jev as the first model in its System One category: unstructured state goes in, and typed, calibrated decisions that software can use directly come out. Jev's architecture and training data are unpublished, but many open typed-decision models have followed, including Nimble, the Decision 1.0-3.0 family and Clef. They answer structured questions about a piece of text, such as a yes/no check, a choice among labelled options or a graded score, by reading the text with a language model and scoring the options with a small readout. Each pairs a different readout with a different backbone, training mix and recipe, so it is unclear which design choices matter. We re-implement five readouts and train all of them under one protocol: the same backbone, the same 15 public English training sources, the same recipe with a per-head learning-rate sweep, and the same seed. Our experiments demonstrate that: The complexity of the head does not affect accuracy: on Qwen3-0.6B every readout reaches 81.3-82.0% in-domain accuracy, and a 0.5M-parameter cosine head, which scores each option by its cosine similarity to the question, matches the 27M-parameter joint schema head adapted from Clef in-domain at half its inference time. Reading all of a text's questions from one prompt also matches one prompt per question in accuracy, with 1.6 times faster inference and 2.1-2.5 times faster training on one H100. The backbone's size matters more than either choice: moving from Qwen3-0.6B to Qwen3-4B adds 4 points in-domain and 11 points on held-out sources, while the base, instruct and thinking checkpoints of the same model tie. We recommend spending effort on the backbone and keeping the head minimal.
- [882] arXiv:2611.16128 (replaced) [pdf, html, other]
-
Title: CoordX Types: A Trustless, Swappable Type System for Shared RecordsComments: 15 pages, 1 image, 8 tables, 1 code blockSubjects: Cryptography and Security (cs.CR)
The CoordX Protocol carries signed and encrypted entries between parties. It proves who wrote each entry and who had permission to write it. It does not say what an entry means. That task belongs above the protocol layer. This paper describes the CoordX type system, the layer that gives entries a meaning. A type is a short document that says how to read one part of a shared record: which entries count, what each position means, what a party may do there, and what a party must do next. The type system has two key properties. It is trustless: a type is an ordinary signed object fixed by its digest, and every party evaluates it for itself over entries that the CoordX Protocol has already verified. No party has to trust a server, a registry or another party's software. It is swappable: a type names the interpreter it is written for, so many interpreters, of different capabilities and designs, can be used side by side in one record. We describe how to write type schemas, and we define the expressive power of the schema language. Payloads form a regular language. Entry structure forms a regular tree language. Every rule and every computed value is a total function in the primitive recursive class, so evaluation always halts. We compare this with KQML, FIPA-ACL and CBCL, and we relate it to language-theoretic security.
- [883] arXiv:2611.16174 (replaced) [pdf, html, other]
-
Title: JudgeMix: Judge-Guided Mixed-Rollout Data Synthesis for LLM AgentsSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Supervised fine-tuning (SFT) on expert trajectories is a standard paradigm for training LLM agents. However, it suffers from a well-known training-inference distribution mismatch, calling for target-aware data synthesis strategies that ground the training data in the target model's native state distribution. To bridge this gap, emerging approaches incorporate the expert either to guide mixed rollouts alongside the target model, or to provide dense step-wise supervision. Nonetheless, existing methods apply expert assistance blind to the target model's inherent behavior: they either disrupt the rollout using sub-optimal random intervention heuristics, or they enforce indiscriminate uniform supervision that unnecessarily overwrites the model's already-correct actions. We introduce JudgeMix, a judge-guided data synthesis framework that dynamically interleaves the target model and expert based on step-wise semantic evaluations. JudgeMix queries expert guidance precisely when the target model struggles, ensuring high trajectory success rates while keeping the state distribution aligned with the target model. It additionally uses the judge evaluation to reweight expert supervision on different steps, preserving model's inherent valid behavior while correcting its errors. On two agentic tool-use benchmarks, AppWorld and AutomationBench, JudgeMix demonstrates consistent superior performance to existing methods, with improvements up to 10 pass@1 points.
- [884] arXiv:2611.16228 (replaced) [pdf, html, other]
-
Title: Indic-Task Gains and Regressions after Reinforcement LearningSutanshu Saksena Raj, Daksh Maan, Avinash Benki, Parijat Chatterjee, Pasunuti Prasanjith, Purusharth MalikComments: Accepted at the GlobalSouthAI Workshop, NeurIPS 2026Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
We run reinforcement learning (RL) with GRPO across four environments, starting from a supervised fine-tuned 30B-A3B mixture-of-experts model covering English and eight Indic languages. We compare the model before and after RL across 34 benchmarks in one evaluation campaign. MATH improves by 36.07 points, and most Indic benchmarks also improve. Yet XQuAD-Indic F1 falls 19.32 points as English answers rise from 8.7% to 91.7%; IndicQA falls 8.64, while English XQuAD gains 11.71 F1. A second RL run from the same supervised initialization adds an Indic environment together with language and repetition rewards. Compared with the released model, it improves both passage-answering and instruction-following benchmarks, along with preserving MATH performance. But it loses ground on reasoning and coding benchmarks. Several training choices changed together, so we cannot isolate the effect of any one of the reward terms. Script compliance remains unresolved. Adding a language reward did not ensure use of the requested script. Passage answering and script compliance therefore need direct evaluation, even when other benchmark scores improve.
- [885] arXiv:2611.16265 (replaced) [pdf, html, other]
-
Title: Rectified Looped TransformerSubjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Looped Transformers increase effective depth by repeatedly applying a shared block, offering a parameter-efficient approach to language modeling. However, most existing works directly reuse each loop's output as the next's input, which can lead to hidden-state scale growth and perturbation amplification as recurrence deepens. We propose Rectified Looped Transformer (RLT), which redesigns feedback between loops using an interpolation schedule inspired by flow-based modeling. At each loop, RLT forms the next input by linearly interpolating between the current state and its prediction. Such interpolation distributes state changes across loops, encouraging a smoother trajectory from the initial input to the final output. Theoretical analysis proves that, under appropriate assumptions, RLT keeps state magnitudes uniformly bounded across loops and reduces the worst-case perturbation amplification. We evaluate RLT on language modeling corpora and downstream benchmarks. At matched training and inference depths, RLT consistently achieves lower perplexity than a standard looped Transformer, while exhibiting greater robustness to input perturbations and variance in inference depth. Consistent with the theoretical results, latent trajectory analysis shows that RLT maintains more stable hidden-state norms across loops, which also facilitates extrapolation-based decoding strategies. Notably, RLT achieves these gains without requiring additional trainable parameters or extra recurrent block evaluations, thus providing a simple yet more stable feedback mechanism for looped Transformers.
- [886] arXiv:2611.16318 (replaced) [pdf, html, other]
-
Title: Benchmarking AI risk behaviorsSubjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
We evaluate LLM risk propensities across a range of contexts and task demands. We evaluate models for how risk averse or risk tolerant they are in various conditions, as well as how stable and consistent their behavior is across conditions. We present a toolkit for developing model risk profiles that can be used to predict model behavior and which can be adapted to include new evaluations of risk propensities. We also examine how these properties have changed across model generations and families. We observe that 2026 state-of-the-art models tend to be more stable, more consistent, and more risk neutral than older models. These trends appear to be related; newer models are more stable and consistent because they are more likely to use expected value maximization as a decision procedure.
- [887] arXiv:2611.16327 (replaced) [pdf, html, other]
-
Title: Designing Information in a Forest: Algorithmic AspectsSubjects: Computer Science and Game Theory (cs.GT); Theoretical Economics (econ.TH)
The Bayesian persuasion framework models strategic information revelation by a sender to one or more receivers. We study an intricate version of the multi-receiver persuasion setting in which the communication hierarchy is a forest - i.e., some receivers observe more informative signals than others, and the partial ordering of the receivers according to the information they get is represented by a pre-specified directed forest. This model naturally captures interactions in hierarchical organizations, and it lies in between the two most commonly studied versions of Bayesian persuasion: private persuasion and public persuasion. Inspired by the classic model of private persuasion, we focus on a setting with a binary state of nature and binary-action heterogeneous receivers s.t. the sender's utility is a monotone function of the set of receivers taking a specific desirable action.
We extend the computational tractability result for symmetric sender's utility from private persuasion to the forest hierarchy framework using methodology inspired by optimal transport theory. We further provide a polynomial-time algorithm when the sender's utility is modular. On the negative side, we show that, unlike in private persuasion, the case of a supermodular sender's utility function leads to an NP-hard optimization problem which further does not admit an FPTAS unless P=NP; this result holds even when each tree component of the forest is a chain of 3 receivers. To complement the hardness result, we provide a simple approximation algorithm with the approximation ratio depending linearly on the depth of the forest hierarchy. We further provide an analogous algorithm for the submodular utility case. Finally, beyond forest communication hierarchies, we rule out an FPTAS for the sender's optimal utility even in the symmetric utility case, unless P=NP. - [888] arXiv:2611.16351 (replaced) [pdf, html, other]
-
Title: HyBridge: Bridge-Guided Self-Supervised Hyperbolic Hypergraph Representation LearningSubjects: Social and Information Networks (cs.SI); Artificial Intelligence (cs.AI)
Hypergraphs capture groupwise relationships among multiple nodes and often exhibit tree-like structures. Although hyperbolic space is well suited for modeling such structures, simply extending existing hypergraph representation learning (HRL) methods to hyperbolic space does not consistently improve performance. In this paper, we investigate the relationship between node structural properties and hyperbolic embedding radii and find that bridge nodes tend to lie closer to the center in graphs, whereas this relationship is not consistently observed in hypergraphs. We further find that conventional bridge score estimation may fail to capture the groupwise connectivity of hypergraphs. Motivated by these findings, we propose HyBridge, a self-supervised hyperbolic HRL framework that (1) estimates node bridge scores by considering the external connectivity of individual hyperedges, (2) learns node and hyperedge embeddings through hyperbolic hypergraph encoding, (3) guides node radial positions according to their bridge scores, and (4) employs contrastive learning to preserve node discriminability. Extensive experiments on 10 real-world hypergraphs demonstrate that HyBridge outperforms 22 competing methods, achieving performance improvements of up to 9.82% in node classification. For reproducibility, we provide the code at this https URL.
- [889] arXiv:2611.16408 (replaced) [pdf, html, other]
-
Title: Gravity and Flight: Learning to Navigate Fine-Grained Remote Sensing ScenesSubjects: Computer Vision and Pattern Recognition (cs.CV)
Vision-language models are increasingly capable, yet their reasoning in complex remote sensing scenes can remain coarse-grained and insufficiently grounded in task-relevant visual evidence. We propose Gravity and Flight (G&F), a framework for scene specialization with limited supervision that separates visual and reasoning verification while coupling policy updates. It combines Evidence-of-Thought supervised fine-tuning (EoT-SFT) and Relational Transfer Policy Optimization (RTPO). EoT-SFT teaches fine-grained observation and faithful reasoning through description, analysis, and supported conclusions. RTPO independently samples responses to scene-matched image pairs with fine-grained visual differences under a shared task. Its rubric-based verification combines Gravity, which checks visual fidelity and description coverage, and Flight, which takes descriptions as premises to assess evidence use and relations between paired judgments. Their scores and task correctness yield a shared pair-level advantage that couples policy updates from both responses. Using 1,008 demonstrations and 500 training pairs for harbor specialization, G&F achieves a mean score of 69.12 across six evaluation sets, exceeding the backbone and supervised initialization by 9.76 and 4.56 points. Matched controls show gains over single-image and joint-input RL, and removing either verification view reduces performance. Applying G&F to urban specialization also improves on the base model and supervised initialization across three evaluation subsets. Project details and updates are available at this https URL.
- [890] arXiv:2611.16423 (replaced) [pdf, html, other]
-
Title: VOKE: Visual Object Kinematics Estimator for Dexterous In-Hand ManipulationSubjects: Robotics (cs.RO)
Vision-based dexterous in-hand manipulation requires knowing not only where an object is, but also how it moves. However, most robotic manipulation systems rely on off-the-shelf visual perception models whose outputs are limited to static object masks, geometry, or 6D poses, leaving downstream controllers to infer object dynamics implicitly from noisy visual observations. We present VOKE, a kinematics-aware perception framework that augments existing object and hand-object pose estimators to recover a complete object kinematic state comprising pose, linear and angular velocity, and linear and angular acceleration. VOKE contains two core perception modules and a downstream application study. First, a Pluggable Kinematics Module (PKM) attaches to the output of a standard pose estimator and predicts the object's linear and angular velocity and acceleration from temporal visual features. Second, a Kinematics-guided Pose Refinement Module (KPRM) converts this motion vector into a second-order, motion-prior reference pose and refines the per-frame pose estimates. We evaluate VOKE on DexYCB and OakInk, where it consistently improves object pose estimation across three different backbones and substantially reduces velocity and acceleration estimation error compared with pose differencing. As a downstream application, we use Kinematics-Augmented Reinforcement Learning (KARL) to compare a kinematics-augmented policy with a pose-history baseline. After training the policies in simulation, we transfer the kinematics-augmented policy to the physical robot without policy fine-tuning, demonstrating the complete perception-to-control pathway in the target domain.
- [891] arXiv:2611.16509 (replaced) [pdf, other]
-
Title: Sound Automatic Lock Placement for Concurrent Programs with PointersComments: 39 pages, 7 figuresSubjects: Logic in Computer Science (cs.LO)
We propose a lightweight method for automatic lock placement in concurrent POSIX programs with explicit pointer manipulation, such as C programs. Given a program that lacks sufficient user-placed mutexes, or contains none at all, our method introduces new mutexes to prevent data races. The method is formally proven to guarantee the absence of data races and mainly consists of three steps. The first step performs lockset analysis and may-happen-in-parallel analysis. The second step conducts a concurrent points-to analysis derived from a Steensgaard-like procedural analysis, where per-thread points-to graphs are unified using a special completion rule. The final step places locks to prevent data races while respecting a lock ordering to avoid deadlocks. In addition, user-annotated critical sections are protected through insertion of dedicated locks. Since lock placement is a complex and error-prone task, our method aims at helping developers write correct concurrent programs.
- [892] arXiv:2611.16525 (replaced) [pdf, html, other]
-
Title: ZONOS3-Mini Technical ReportSubjects: Sound (cs.SD); Artificial Intelligence (cs.AI)
ZONOS3-Mini is an ultra-compact, ultra-fast, zero-shot voice-cloning TTS model. The model can run in real time on a laptop or smartphone. For its latency and size, it achieves state-of-the-art accuracy and cloning fidelity. To achieve this we propose a latent MoE backbone with only 886M parameters (100M active per token) and a new audio codec, Zyphra Audio Codec 1 (ZAC-1), which represents 44.1 kHz audio at a 15 Hz frame rate. On the new ZTTS2-Eval benchmark, ZONOS3-Mini reaches a 6.71% macro-average WER over 55 language sets (35 clean and 20 in the wild), which is competitive with larger and slower state-of-the-art models. We release the model weights, inference, and training code.
- [893] arXiv:2611.16803 (replaced) [pdf, html, other]
-
Title: Detection Is Not Early Warning: Evaluating Token-Level Alarms on Language Model FailuresMarco De Negri, Luca Sartori, Lorenzo Baggi, Moritz Reihs, Eduard Frank Ďurech, Mennatallah El-Assady, Valentina Pyatkin, Anna HedströmComments: Accepted at the NeurIPS 2026 Workshop on Interpretability as a Science. 30 pages (10 main text), 6 figures, 16 tables. Code: this https URLSubjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Many failures of language models begin at a specific token, the onset. A monitor that reads the model during generation can classify the finished response, detect the failure once it is under way, or warn before it begins, and only a warning leaves time to stop or steer the generation. Monitors are usually evaluated with response-level metrics such as AUROC, which ignore when the monitor fires. We introduce PREWARN, an evaluation protocol for any failure with an annotated onset. It calibrates each monitor's threshold on validation data to an error rate chosen for the application, runs the monitor, and scores its alarms relative to the onset on all three tasks. As a case study we apply it to repetitive degeneration in three open language models, with early warning as the main task, using linear probes on activations and six untrained baselines. Response-level metrics saturate on this failure, and a rule that reads only the response length scores almost perfectly, while probes that AUROC ranks as equal warn very differently under PREWARN. Probes warn before the onset on most degenerate responses of two of the three models, low-rank adapters below the probe more than double the share of tokens just before the onset that it flags, largely by flagging degenerate responses long before the onset, and no untrained baseline warns much more often than random alarms. We release the probes with calibrated thresholds at this https URL
- [894] arXiv:2611.16931 (replaced) [pdf, other]
-
Title: Beyond Curiosity: Habit and Organizational Learning Capability as Predictors of Shadow AIComments: Author accepted manuscript; v1 is the submitted version. Received 10 June 2026; revised 14 October 2026; accepted 31 October 2026. Published in Computers in Human Behavior, Article 109243. (c) 2026. This manuscript version is made available under the CC-BY-NC-ND 4.0 license. 58 pages including supplementary materialsJournal-ref: Computers in Human Behavior (2026), Article 109243Subjects: Computers and Society (cs.CY); Human-Computer Interaction (cs.HC)
Generative AI has entered the workplace faster than organizations can govern it. Employees may use AI tools their employer has not sanctioned and conceal that use, but the psychology of this behavior remains poorly understood. We compare an exploratory-disposition account with one centered on reported use and habit. Using survey data from 510 employed adults reporting workplace generative-AI use during the preceding 12 months, we estimated the unique and relative contributions of workplace curiosity, habit, technology-acceptance beliefs, and perceived organizational learning capability through hierarchical regression, bootstrap relative weights, and full-subset dominance analysis. Habit had the largest relative weight, accounting for 28.3% of explained variance, and completely dominated joyous exploration across all 512 possible predictor subsets. Joyous exploration contributed 3.4% and became nonsignificant when use behavior, performance expectancy, and social influence entered the model; habit subsequently explained additional variance. Organizational learning subscales accounted for 24.1% of explained variance; experimentation and knowledge transfer were associated with lower shadow AI, while curiosity-by-learning interactions were nonsignificant. Separate models for unauthorized use and concealment reproduced the habit-dominance pattern. Perceived professional support was associated with lower concealment, although its coefficient did not significantly differ between the two behavioral outcomes. Shadow AI in this sample was more strongly associated with self-reported habit strength and perceived organizational conditions than with exploratory disposition. The cross-sectional design does not establish causal or temporal order.
- [895] arXiv:2611.16934 (replaced) [pdf, html, other]
-
Title: Time Flies When You Parallelise: Pareto-efficient Time Series Foundation Models with Linear RecurrencesSubjects: Machine Learning (cs.LG)
Time series foundation models have demonstrated strong zero-shot forecasting performance, but computational efficiency remains a central architectural challenge. Temporal self-attention incurs quadratic cost in the number of temporal tokens, while nonlinear recurrent state transitions restrict parallel computation along the temporal dimension. We introduce ci-proto, a family of time series foundation models based on Mamba-3, whose linear state recurrences support chunk-wise parallel-in-time computation during training, and inference costs that scale linearly with sequence length. The architecture alternates recurrent temporal processing with cross-variate attention, supporting multivariate probabilistic forecasting with past-only and future-known covariates. Across fev-bench's 100 forecasting tasks, ci-proto XL sets a new state of the art among time series foundation models. The five ci-proto variants also make up the entire Pareto frontier for mean scaled quantile loss and mean absolute scaled error against both parameter count and median runtime.
- [896] arXiv:2611.16937 (replaced) [pdf, html, other]
-
Title: S1-Evo: Experience-Guided Evidence Competition Attacks on Typed System-1 Decision ModelsComments: noneSubjects: Artificial Intelligence (cs.AI)
Typed System-1 (S1) decision models are increasingly deployed in LLM agents for routing, moderation, and authorization to cut inference cost, yet untrusted context can still manipulate typed outputs and alter execution branches. Unlike autoregressive System-2 (S2) models, S1 interfaces score dynamic rules and options without token-by-token decoding, mismatching attacks targeting text continuation or explicit override commands. We propose S1-Evo, a self-evolving framework tailored to S1's contract-based option competition and margin feedback. S1-Evo perturbs untrusted fields via contract-aware evidence channels to induce \emph{evidence competition} with original task evidence while preserving fact invariance. Offline on seed tasks, it optimizes candidates via feedback search guided by directional margins and re-verification, distilling verified flips into experience cards. At test time, S1-Evo retrieves relevant cards to construct perturbations for unseen tasks. Across four benchmarks and three S1 models, S1-Evo reaches 71.99\% mean ASR and 57.47\% high-confidence ASR on \texttt{jev-1.14}, and causes up to 24.00\% E2E agent violations.
- [897] arXiv:2611.17108 (replaced) [pdf, html, other]
-
Title: LazyWeights: Decode-on-Demand Edge LLMsComments: 6 pages, 1 reference page, 5 figuresSubjects: Machine Learning (cs.LG); Hardware Architecture (cs.AR); Computation and Language (cs.CL); Information Theory (cs.IT)
Memory decides which LLMs an edge device runs, and how fast: capacity on a 4 GB Jetson Nano, bandwidth on an 8 GB Jetson Orin Nano Super. Deployed formats (GPTQ, AWQ, this http URL's K- and I-quants) spend equal bits on every integer code of a matrix, although the codes are far from uniform. LazyWeights keeps these formats but tunes GPTQ so the codes compress, Huffman-codes them losslessly in interleaved lanes, and decodes them on demand in a small unit between DRAM and L2, so GPU kernels run unchanged and DRAM moves only coded bytes. At 1% and 3.5% perplexity budgets, coded Phi-3-mini, Mistral-7B and SmolLM-1.7B are 5.7-31.4% smaller than the best this http URL format. A bit-exact 0.21 mm^2 ASIC decoder unit (1 GHz, 7 nm) cuts Orin latency and energy per token by up to 36% and 38% (54% and 44% with prompt-lookup speculation). On the Nano, LazyWeights needs 58-64% less time and 53-57% less energy per token than CPU this http URL, and runs a streamed Mistral-7B 6.6x faster at 3.89 bits per weight. A Huffman decoder beats an rANS decoder in rate and area at equal throughput; the ASIC decoder unit runs bit-exact on a Zynq UltraScale+ FPGA (AMD ZCU102).
- [898] arXiv:2611.17143 (replaced) [pdf, other]
-
Title: A Convex-Analytic Reformulation for Poisoning Attacks on Logistic RegressionSubjects: Machine Learning (cs.LG); Optimization and Control (math.OC)
We study data poisoning attacks on logistic regression with $\ell_2$-regularization, where an attacker chooses the features of a fixed number of injected observations with prescribed labels to maximize the logistic loss on the clean training data. Strong convexity of the defender's problem yields an exact reformulation through its first-order optimality conditions. Using convex conjugacy of the softplus function, we replace the sigmoid relations by a single convex optimality-gap inequality and express the attacker's objective as a concave function. A change of variables makes the stationarity condition affine, leaving bounded bilinear equalities as the only nonconvex constraints. We derive finite bounds on all variables involved in these equalities, including a global bound on the coefficient norm and tighter bounds obtained by preprocessing based on convex optimization. Experiments on eight binary classification datasets evaluate attack effectiveness and compare the two reformulations when solved with SCIP. Attacks obtained by local optimization degrade predictive performance on most datasets, while the convex-analytic formulation produces substantially stronger reported dual bounds within a fixed time limit.
- [899] arXiv:2611.17148 (replaced) [pdf, html, other]
-
Title: Learning Dynamics of the Influencer's Game in Resource LandscapesComments: 23 pages 6 figuresSubjects: Computer Science and Game Theory (cs.GT); Theoretical Economics (econ.TH); Physics and Society (physics.soc-ph)
Across domains as varied as ecology, economics, and online platforms, individuals are incentivized to position themselves within resource-rich strategy spaces. In this work we introduce a new spatial game framework called the influencer's game. Within this new framework players' strategies are represented as positions within a trait space. Players then receive a share of a portion of a resource based on their relative influence at that resource's points traits. For the influencer's game, we show that there is a unique symmetric Nash equilibrium that exists under strict log-concavity of the influence kernel. Using the unique equilibrium we show that the stability of the system depends on the influence kernel's parameters, number of players, and the resource distribution shape. For any dimension we demonstrate a stability threshold that depends on both the scatter and curvature of the players. When this threshold is crossed, the symmetric equilibrium destabilizes and gives way to a variety of bifurcation types; including, specialist clusters, limit cycles, and sub-games that reproduce the larger game's structure. We ground these results using real-world data for resources as a demonstration of the game's application. Finally we extend the game dynamics to use multi-agent reinforcement learning and further show that the predicted bifurcation structure persists under partial observability and discretized strategy spaces. In all, the influencer's game provides a novel framework for predicting when competitive pressure drives convergence versus fragmentation with possible applications across domains.
- [900] arXiv:2611.17238 (replaced) [pdf, html, other]
-
Title: HEIR: Harness Egocentric Intent for Human-Robot InteractionsYichi Zhang, Bo Wang, Zhenghan Zhu, Yutong Han, Jiayin Zou, Gao Kejun, Chang Liu, Shaoxuan Xie, Chi Zhang, Shi Haoyang, Zongxuan Lai, Sen Cui, Guocai Yao, Miao LiuSubjects: Robotics (cs.RO)
For robots to assist, serve, and collaborate with humans, they must not only interpret human requests conveyed through language and multimodal intentional cues, but also execute long-horizon tasks that interleave navigation and manipulation. In this paper, we introduce HEIR, a resource for studying how to Harness Egocentric Intent for Human-Robot Interaction. HEIR pairs human speech, measured gaze, and egocentric video with robot observations and teleoperated navigation-manipulation trajectories, while each session preserves successive episodes, with each request fulfilled through multiple execution subtasks. The captured data, together with human-authored annotations, support the study of intent grounding, execution planning, and task completion within a unified interaction setting. To demonstrate the value of HEIR, we conduct extensive offline experiments to evaluate how models benefit from additional human-centric context across these tasks, together with real-robot experiments that assess execution task completion. HEIR thus provides a valuable resource for studying user-centric robot interaction in naturalistic household environments. Our demo and dataset are available online.
- [901] arXiv:2611.17241 (replaced) [pdf, html, other]
-
Title: Ballad: Bandit-Based LLM Routing for Automated Heuristic DiscoverySubjects: Artificial Intelligence (cs.AI)
Modern automated heuristic discovery (AHD) systems increasingly rival hand-crafted heuristics by leveraging Large Language Models (LLMs) to iteratively mutate and refine solutions. A natural assumption is that more powerful and more expensive models consistently yield better heuristics. We challenge this empirically: across multiple NP-hard combinatorial optimization problems, we find no consistent advantage for larger, costlier models. This raises a deeper question: given a pool of LLMs spanning frontier and freely hosted or open-weight models, how should one adaptively select which model to invoke at each step of the search? The challenge is non-trivial. The optimal LLM may shift over the course of a run, and because one LLM's output seeds the context for subsequent calls, each model's reward signal is non-stationary and entangled across the heuristic population, thus violating the independence assumptions of classical online decision-making settings. We address this through Ballad (Bandit-based LLM Routing for Automated Heuristic Discovery), an online bandit algorithm that learns to orchestrate a pool of LLMs into an ensemble that matches or exceeds what any soloist LLM achieves alone. This orchestration is powered by an LLM selection policy that favors rare, high-scoring heuristics over stable averages, a DAG-based credit propagation mechanism through the genealogy of the heuristic population, and discounting of stale observations to remain attuned to each model's current utility. Across five NP-hard combinatorial optimization problems and a pool of four models of varying capability and cost, Ballad not only matches the best individual LLM in hindsight but, in several cases, surpasses every individual model while reducing API cost by up to 65.6% relative to always using the strongest model.
- [902] arXiv:2611.17332 (replaced) [pdf, html, other]
-
Title: ArchSpace: Recording, Reviewing, and Referencing LLM Architecture ResearchJie Wang, Tong Zhu, Jiahao Kuang, Zhe Xu, Mowen Ruan, Yunkai Chai, Yuliang Liu, Dehong Zheng, Rubin Wei, Yu Cheng, Bowen Zhou, Tao Ji, Zhouhan Lin, Qi ZhangSubjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Model architecture shapes the scaling behavior, training efficiency, and inference cost of LLMs, making it a central focus of both industry and academia. In industry, teams such as those behind DeepSeek, Qwen, and Kimi routinely introduce architectural innovations, yet public reports rarely document which ideas succeeded, which were abandoned, and why. In academia, limited compute and inconsistent baselines often restrict evaluations to small-scale, single-point comparisons, making it difficult to establish whether improvements hold at larger scales or translate into broader impact. ArchSpace addresses both gaps as a public platform for Recording, Reviewing, and Referencing Architecture Research: each proposal is recorded with its experiments and reasoning, reviewed by the community, and referenced by later work. To evaluate ideas before costly large-scale training, ArchSpace leverages scaling laws to predict model loss at larger scales from smaller experiments, a practice widely adopted in industry. We provide initial references that make the platform usable: Qwen3.5 and OLMo3 baselines, fitted scaling laws, and a reusable training recipe. Ideas validated through this review can in turn become new baselines, and ArchSpace thereby aims to broaden the range of ideas that shape the next generation of models.
- [903] arXiv:2211.03997 (replaced) [pdf, html, other]
-
Title: Online Nonconvex Decision Making with Approximate OraclesSubjects: Optimization and Control (math.OC); Data Structures and Algorithms (cs.DS)
We study online decision making with possibly nonconvex (for example, combinatorial or integer) local feasible sets and convex long-term constraints. At each period, a decision maker selects a local action to maximize cumulative reward while ensuring that the sequence of actions satisfies the global constraints in the long term. It is known that $O(\sqrt{T})$ long-term constraint violation and $O(\sqrt{T})$ reward regret are achievable if given access to a local optimization oracle that exactly solves optimization problems over each local feasible set. Instead of requiring exact optimization over each local feasible set, we assume access to an approximate local optimization oracle that permits optimization errors quantified by two parameters $(\delta_+,\delta_\times)$. We develop an approximate online dual mirror descent algorithm and show that its guarantees depend on a feasibility slack parameter $\underline d$, which measures the compatibility between local feasible sets and the long-term constraint. Under the standard random permutation model, when $\delta_+=O(1)$, $\delta_\times<\underline d$ and $\underline d-\delta_\times=\Omega(1)$, the algorithm achieves $O(\sqrt T)$ long-term constraint violation and reward regret up to $O\bigl(\sqrt T+\delta_\times T\bigr)+\delta_+ T$. We also show a matching limitation: when $\delta_+=0$ and $\delta_\times>\underline d$, no online algorithm relying on such approximate oracle outputs can guarantee sublinear long-term constraint violation in general.
- [904] arXiv:2409.12566 (replaced) [pdf, html, other]
-
Title: Quantum Channel Testing in Average-Case DistanceSubjects: Quantum Physics (quant-ph); Computational Complexity (cs.CC); Data Structures and Algorithms (cs.DS)
We study the complexity of testing properties of quantum channels. First, we show that testing identity to any channel $\mathcal N: \mathbb C^{d_{\mathrm{in}} \times d_{\mathrm{in}}} \to \mathbb C^{d_{\mathrm{out}} \times d_{\mathrm{out}}}$ in diamond norm distance requires $\Omega(\sqrt{d_{\mathrm{in}}} / \varepsilon)$ queries, even in the strongest algorithmic model that admits ancillae, coherence, and adaptivity. This is due to the worst-case nature of the distance induced by the diamond norm.
Motivated by this limitation and other theoretical and practical applications, we introduce an average-case analogue of the diamond norm, which we call the average-case imitation diamond (ACID) norm. In the weakest algorithmic model without ancillae, coherence, or adaptivity, we prove that testing identity to certain types of channels in ACID distance can be done with complexity independent of the dimensions of the channel, while for other types of channels the complexity depends on both the input and output dimensions. Building on previous work, we also show that identity to any fixed channel can be tested with $\tilde O(d_{\mathrm{in}} d_{\mathrm{out}}^{3/2} / \varepsilon^2)$ queries in ACID distance and $\tilde O(d_{\mathrm{in}}^2 d_{\mathrm{out}}^{3/2} / \varepsilon^2)$ queries in diamond distance in this model. Finally, we prove tight bounds on the complexity of channel tomography in ACID distance. - [905] arXiv:2501.04151 (replaced) [pdf, html, other]
-
Title: Efficient LP warmstarting for linear modifications of the constraint matrixSubjects: Optimization and Control (math.OC); Computational Complexity (cs.CC)
We consider the problem of computing the optimal solution and objective value of a linear program whose constraint matrix changes linearly with a parameter, for a predefined set of values of this parameter. Based on the information given by a precomputed basis, we present three efficient LP warm-starting algorithms, based on the eigenvalue decomposition, the Schur decomposition and the Hessenberg decomposition, to evaluate the optimal solution and optimal objective value of these problems. The three algorithms have an overall complexity of $O(pm^2+pmn)$ where $m$ (resp. $n$) is the number of constraints (resp. variables) of the original problem and $p$ the number of parameter values after an initial preprocessing step. We also provide theorems related to the optimality conditions to verify when a basis is still optimal. When the perturbation of the basis matrix has a low rank $\rho$, as is often the case in practice, we then reduce all the computations to a $\rho\times\rho$ matrix built with $\rho$ solves with the factorization of the basis maintained by the simplex solver; with it, we compute, in each direction, an interval on which the basis is guaranteed to remain optimal, together with a bound on the objective. In practice, these intervals are close to the exact ones, and they remain useful for degenerate bases. We also propose a matrix-free method based on shifted Krylov subspaces. Numerical experiments on 398 instances built from real models, with up to 349,120 constraints, assess the accuracy, the cost and the range of applicability of all the methods. On the interval on which a basis remains optimal, once this interval is known, the low-rank and Krylov methods compute the optimal objective value 117 to 304 times faster than a warm start of the solver, in median.
- [906] arXiv:2505.17965 (replaced) [pdf, html, other]
-
Title: Bias-Optimal Bounds for Stochastic Gradient DescentComments: 30 pages. Under reviewSubjects: Optimization and Control (math.OC); Machine Learning (cs.LG); Machine Learning (stat.ML)
The non-asymptotic analysis of Stochastic Gradient Descent (SGD) typically yields bounds that decompose into a bias term and a variance term. In this work, we focus on the bias component and study the extent to which SGD can match the optimal convergence behavior of deterministic gradient descent. Assuming only (strong) convexity and smoothness of the objective, we derive new bounds for the full range of constant step-sizes $\gamma L \in (0,2)$. This includes critical and large step-size regimes that were previously unexplored without additional variance assumptions, in both the convex and strongly convex cases. Those bounds are bias-optimal, in the sense that the bias term coincides with the worst-case rate of gradient descent, for all non-critical step-sizes. For critical step-sizes, we show that no bounds can be obtained with an optimal bias term, but that this optimal bias can be approximated arbitrarily closely. As an immediate application, we provide new complexity bounds for the Stochastic Proximal algorithm, without the need of regularity assumptions.
- [907] arXiv:2507.18520 (replaced) [pdf, html, other]
-
Title: Recovering the Euclidean Geometry of High-Dimensional Data Under Heterogeneous NoiseSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST)
Pairwise Euclidean distances provide a simple geometric representation of high-dimensional data and underlie many data analysis methods, including clustering, visualization, and manifold learning. However, in real-world applications, these distances can be severely corrupted by heterogeneous noise, where different data points are subject to different, and potentially large, noise levels. Such noise distorts the distances inhomogeneously, leading to a misrepresentation of the underlying data geometry. In this work, we ask whether the underlying clean Euclidean geometry can nevertheless be recovered from the corrupted pairwise distances. We answer this question affirmatively, showing that, perhaps surprisingly, such recovery is possible across a broad class of high-dimensional settings with underlying low-dimensional geometry, even when the noise magnitudes vary substantially across observations. We develop a principled approach for recovering the clean pairwise distances, without prior knowledge of either the clean data structure or the noise distribution, and establish probabilistic recovery guarantees. In particular, the recovery error converges to zero at polynomial rates as the feature dimension and dataset size increase simultaneously. Experiments on synthetic datasets demonstrate accurate recovery of the Euclidean geometry in challenging regimes and improved robustness of downstream methods that rely on pairwise distances. When applied to real single-cell RNA sequencing data, our method reliably recovers local neighborhood structure, enabling accurate nearest-neighbor identification used in many downstream analyses.
- [908] arXiv:2509.05894 (replaced) [pdf, html, other]
-
Title: Toric geometry of ReLU neural networksSubjects: Algebraic Geometry (math.AG); Machine Learning (cs.LG); Machine Learning (stat.ML)
Given a continuous piecewise linear function $f:\mathbb{R}^{n_0} \to \mathbb{R}$ and a fixed architecture of feedforward ReLU neural networks, the exact function realization problem asks when a network of that architecture realizes $f$. In contrast, expressivity concerns characterizing all functions exactly realizable by a given architecture. To address these questions systematically, we establish a connection between toric geometry and ReLU neural networks, allowing tools from algebraic geometry to analyze the structures. Starting with an unbiased ReLU network with rational weights, we define the ReLU fan, the ReLU toric variety, and the ReLU Cartier divisor associated with the network. This work also relates the tropical and the toric geometry: the former captures combinatorial data of ReLU networks, while the latter recovers the underlying geometric structure. As an application, we obtain a complete characterization of the functions realizable by unbiased shallow ReLU networks with rational weights. The characterization is expressed in terms of intersection numbers of ReLU Cartier divisors with torus-invariant curves and leads to a decomposition of piecewise-linear functions into shallow and non-shallow components.
- [909] arXiv:2512.05024 (replaced) [pdf, html, other]
-
Title: Model-Free Assessment of Simulator Fidelity via Quantile CurvesComments: Accepted to NeurIPS 2026Subjects: Methodology (stat.ME); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
As generative AI models are increasingly used to simulate real-world systems, quantifying the ``sim-to-real'' gap is critical. For each input setting of interest -- which we call a \emph{scenario}, such as a survey question or operating condition -- the real and simulated systems are associated with unobserved latent population parameters, and their discrepancy varies across scenarios. A fundamental challenge is that, for any given scenario, this discrepancy cannot be observed directly, since both systems are accessible only through finite samples, often of heterogeneous sizes across scenarios. Standard predictive inference methods are therefore ill-suited, as they quantify uncertainty in observable outputs rather than latent population parameters. To address this, we construct confidence sets for these latent parameters and use them to derive a robust proxy for the sim-to-real discrepancy. We then estimate the quantile function of this proxy to obtain a distribution-level risk profile of the simulator, which supports a broad range of statistical summaries, including statistical inference for the real output distribution in a new scenario, the calculation of risk measures like Conditional Value-at-Risk (CVaR), and principled comparisons across simulators. Our method is model-agnostic and handles general output spaces, such as categorical survey responses and continuous multi-dimensional data. We demonstrate the practical utility of this method by evaluating the alignment of four major LLMs with human populations on the WorldValueBench dataset.
- [910] arXiv:2601.03064 (replaced) [pdf, html, other]
-
Title: Similarity-Sensitive Entropy under Representation Change and InferenceComments: 26 pages. v3: Revised abstract; generalized coarse graining proof; added further directions; revised theorem statementsSubjects: Probability (math.PR); Information Theory (cs.IT)
Similarity-sensitive entropy measures the uncertainty of a probability law relative to a similarity kernel that encodes the distinguishability of states. We develop the measure-theoretic treatment covering both finite similarity matrices and general probability spaces, and study how the law and similarity kernel transform under measurable maps, channels, and conditioning operations. This yields data-processing inequalities, so a reduction in entropy quantifies how much distinguishability is lost under representation change. We also define a conditional similarity-sensitive entropy theory, give a counterexample to a recent conjecture on concavity, and identify a practical one-dimensional Laplace kernel class where concavity holds.
- [911] arXiv:2601.21843 (replaced) [pdf, html, other]
-
Title: The Leibniz adjunction in homotopy type theory, with an application to simplicial type theoryComments: v4: Corrected links to Agda HTML. To appear at CSL 2027Subjects: Category Theory (math.CT); Logic in Computer Science (cs.LO); Logic (math.LO)
Simplicial type theory extends homotopy type theory and equips types with a notion of directed morphisms. A Segal type is defined to be a type in which these directed morphisms can be composed. We show that all higher coherences can be stated and derived if simplicial type theory is taken to be homotopy type theory with a postulated interval type. In technical terms, this means that if a type has unique fillers for $(2,1)$-horns, it has unique fillers for all inner $(n,k)$-horns. This generalizes a result of Riehl and Shulman for the case $n = 3, k \in \{1, 2\}$. Our main technical tool is the Leibniz adjunction: the pushout-product is left adjoint to the pullback-hom in the wild category of types. While this adjunction is well known for ordinary categories, it is much more involved for higher categories, and the fact that it can be proved for the wild category of types (a higher category without stated higher coherences) is non-trivial. We make profitable use of the equivalence between the wild category of maps and that of families. We have formalized the results in Cubical Agda.
- [912] arXiv:2602.16265 (replaced) [pdf, other]
-
Title: On sparsity, extremal structure, and monotonicity properties of Wasserstein and Gromov-Wasserstein optimal transport plansTitouan Vayer (COMPACT)Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
This note gives a self-contained overview of some key properties of the Gromov-Wasserstein (GW) distance, compared with the standard linear optimal transport (OT) framework. More specifically, I explore the following questions: are GW optimal transport plans sparse? under what conditions does there exist an optimal plan supported on a permutation? do optimal plans satisfy a form of cyclical monotonicity? In particular, I present the conditionally negative/positive definite (CND/CPD) properties and show that, when they hold, there exist GW optimal plans that are sparse and supported on a permutation. In the last section, I also give examples where no permutation is optimal when the CND or CPD assumption is dropped.
- [913] arXiv:2603.06740 (replaced) [pdf, html, other]
-
Title: ViroGym: Realistic Large-Scale Benchmarks for Evaluating Viral ProteinsSubjects: Quantitative Methods (q-bio.QM); Artificial Intelligence (cs.AI)
Protein language models (pLMs) have shown strong potential for zero-shot prediction of missense variant effects, yet systematic benchmarking on viral proteins remains limited, a critical gap given the need for proactive tools that can anticipate emerging mutations ahead of experimental validation. Here we introduce ViroGym, a comprehensive benchmark evaluating pLMs across three tasks: 79 deep mutational scanning (DMS) assays covering eukaryotic viruses with 552,065 mutated sequences across 7 phenotypic readouts, 21 influenza neutralisation tasks, and a real-world pandemic prediction task for SARS-CoV-2. We benchmark well-established pLMs on fitness landscapes, antigenic diversity, and pandemic forecasting, and find that the ProGen2 family consistently achieves the strongest performance across all three tasks. Crucially, DMS and neutralisation performance reliably identifies models that generalise to real-world emergence, even though the mutation sets they surface barely overlap, revealing that complementary in vitro benchmarks capture the evolutionary constraints needed for real-world mutation forecasting. The data and code supporting this study are publicly available on GitHub at this https URL
- [914] arXiv:2603.09600 (replaced) [pdf, html, other]
-
Title: A Variational Latent Equilibrium for Learning in Neuronal CircuitsSubjects: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE); Systems and Control (eess.SY); Biological Physics (physics.bio-ph)
Brains remain unrivaled in their ability to recognize and generate complex spatiotemporal patterns. While AI is able to reproduce some of these capabilities, deep learning algorithms remain largely at odds with our current understanding of brain circuitry and dynamics. This is prominently the case for backpropagation through time (BPTT), the go-to algorithm for learning complex temporal dependencies. In this work we propose a general formalism to approximate BPTT in a controlled, biologically plausible manner. Our approach builds on, unifies and extends several previous approaches to local, time-continuous, phase-free spatiotemporal credit assignment based on principles of energy conservation and extremal action. Our starting point is a prospective energy function of neuronal states, from which we calculate real-time error dynamics for time-continuous neuronal networks. With this, as a first contribution we provide a simple and straightforward derivation of the adjoint method dynamical equations, the time-continuous equivalent to BPTT. Importantly, additional approximations are necessary to avoid unphysical temporal non-localities. One such approximation was recently proposed in Generalized Latent Equilibrium (GLE), but enforcing temporal locality of the backpropagating dynamics induces an error in the gain that becomes problematic when processing complex signals. Thus, our second contribution is to design a learning rule for the backward weights to find a fully local (in space and time) set of equations for neuron and synapse dynamics that corrects for this gain difference. With this inclusion of backward weight learning, our theory provides a rigorous framework for spatiotemporal deep learning in the brain, while simultaneously advancing a blueprint for physical circuits capable of carrying out these computations.
- [915] arXiv:2603.20891 (replaced) [pdf, html, other]
-
Title: Auto-differentiable data assimilation: Co-learning of states, dynamics, and filtering algorithmsSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Signal Processing (eess.SP); Dynamical Systems (math.DS)
Data assimilation algorithms estimate the state of a dynamical system from partial observations, where the successful performance of these algorithms hinges on costly parameter tuning and on employing an accurate model for the dynamics. This paper introduces a framework for jointly learning the state, dynamics, and parameters of filtering algorithms in data assimilation through a process we refer to as auto-differentiable filtering. The framework leverages a theoretically motivated loss function that enables learning from partial, noisy observations via gradient-based optimization using auto-differentiation. We further demonstrate how several well-known data assimilation methods can be learned or tuned within this framework. To underscore the versatility of auto-differentiable filtering, we perform experiments on dynamical systems spanning multiple scientific domains, such as the Clohessy-Wiltshire equations from aerospace engineering, the Lorenz-96 system from atmospheric science, the generalized Lotka-Volterra equations from systems biology, and the two-layer quasi-geostrophic model from geophysical fluid dynamics. Finally, we provide guidelines for practitioners to customize our framework according to their observation model, accuracy requirements, and computational budget.
- [916] arXiv:2604.02058 (replaced) [pdf, html, other]
-
Title: Faster Symmetric Rendezvous on Four or More LocationsSubjects: Optimization and Control (math.OC); Computer Science and Game Theory (cs.GT)
In the symmetric rendezvous problem, two players follow the same (randomized) strategy to visit one of $n$ locations in each time step $t=0,1,2,\dots$. Their goal is to minimize the expected time until they visit the same location and thus meet. A canonical strategy due to Anderson and Weber (J. Appl. Probab., 1990) is known to be optimal for $n=2$ and $n=3$ (Weber, Math. Oper. Res. 2012), but whether it remains optimal for larger values of $n$ has been an open question since 1990. We show that it does not remain optimal: for any finite $n\geq 4$, we construct an explicit symmetric strategy that achieves a strictly smaller expected meeting time than the Anderson-Weber strategy.
In the Anderson-Weber strategy players stay at a dedicated home location for $n-1$ steps with a certain probability $\theta$ and with the remaining probability tour all non-home locations in a random order. Our improving strategy introduces carefully chosen correlations between consecutive tours of the non-home locations. The construction is uniform in $n$ and is guided by a graph-theoretic view in which tours correspond to permutations and meetings to edges in the complement of the derangement graph. By exploiting the clique structure of this graph we obtain a correlated strategy that improves on the Anderson-Weber strategy. For $n=4$, we give an exact expression for the improvement; for any $n\geq 5$, we obtain a lower bound on the expected improvement of $\frac{483(1-\theta)^6}{(n-1)^8}$, where $\theta \in [0,1)$ is the probability of staying at the home location. The graph-theoretic framework we introduce may be useful more widely in the design and analysis of correlated strategies for rendezvous. - [917] arXiv:2604.11578 (replaced) [pdf, html, other]
-
Title: Minimizing classical resources in variational measurement-based quantum computation for generative modelingComments: 15 pagesSubjects: Quantum Physics (quant-ph); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML)
Measurement-based quantum computation (MBQC) is a framework for quantum information processing in which a computational task is carried out through one-qubit measurements on a highly entangled resource state. Due to the indeterminacy of the outcomes of a quantum measurement, the random outcomes of these operations, if not corrected, yield a variational quantum channel family. Traditionally, this randomness is corrected through classical processing in order to ensure deterministic unitary computations. Recently, variational measurement-based quantum computation (VMBQC) has been introduced to exploit this measurement-induced randomness to gain an advantage in generative modeling. A limitation of this approach is that the corresponding channel model has twice as many parameters compared to the unitary model, scaling as $N \times D$, where $N$ is the number of logical qubits (width) and $D$ is the depth of the VMBQC model. This can often make optimization more difficult and may lead to poorly trainable models. In this paper, we present a restricted VMBQC model that extends the unitary setting to a channel-based one using only a single additional trainable parameter. We show, both numerically and algebraically, that this minimal extension is sufficient to generate probability distributions that cannot be learned by the corresponding unitary model.
- [918] arXiv:2604.18152 (replaced) [pdf, other]
-
Title: mlr3torch: A Deep Learning Framework in R based on mlr3 and torchSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Deep learning (DL) has become a cornerstone of modern machine learning (ML) praxis. We introduce the R package mlr3torch, which is an extensible DL framework for the mlr3 ecosystem. It is built upon the torch package, and simplifies the definition, training, and evaluation of neural networks for both tabular data and generic tensors (e.g., images) for classification and regression. The package implements predefined architectures, and torch models can easily be converted to mlr3 learners. It also allows users to define neural networks as graphs. This representation is based on the graph language defined in mlr3pipelines and allows users to define the entire modeling workflow, including preprocessing, data augmentation, and network architecture, in a single graph. Through its integration into the mlr3 ecosystem, the package allows for convenient resampling, benchmarking, preprocessing, and more. We explain the package's design and features and show how to customize and extend it to new problems. Furthermore, we demonstrate the package's capabilities using three use cases, namely hyperparameter tuning, fine-tuning, and defining architectures for multimodal data. Finally, we present some runtime benchmarks.
- [919] arXiv:2605.29371 (replaced) [pdf, html, other]
-
Title: Kernel-based potential mean-field games with unbiased random Fourier $U$-statisticsSubjects: Optimization and Control (math.OC); Machine Learning (cs.LG); Numerical Analysis (math.NA); Machine Learning (stat.ML)
We study the subclass of potential mean-field games in which the running interaction cost and the terminal target cost are both expressed through reproducing-kernel maximum mean discrepancy (MMD) penalties, and develop a computational framework that exploits this kernel structure. Both costs are estimated from finite-sample empirical distributions using a random Fourier U-statistic representation that is unbiased and has linear cost in the batch size. The drift of the controlled diffusion is parametrized by a neural network and trained via stochastic gradient descent. For population near-minimizers we prove convergence to the terminal-constrained problem as the penalty diverges, and show that the same limit is recovered almost surely when such controls are evaluated on independent finite samples under explicit coupling conditions on the penalty, random-feature count and sample size. The framework includes the kernel-MMD-penalty Schr{ö}dinger bridge problem as the special case of a vanishing interaction cost. Numerical experiments illustrate the method on the Schr{ö}dinger bridge problem in dimensions up to one hundred, and on an electric vehicle charging coordination problem with per-vehicle physical heterogeneity, where a control-dependent aggregate-demand cost produces a trade-off between peak normalized charging power and the terminal service level.
- [920] arXiv:2606.29553 (replaced) [pdf, html, other]
-
Title: An annihilation-number Caro-Wei bound: a TxGraffiti conjecture and an independence-number bracketComments: 9 pagesSubjects: Combinatorics (math.CO); Discrete Mathematics (cs.DM)
Computable independence bounds give intervals that can be compared by containment or by their worst-case endpoint ratios. For positive maximum degree $\Delta$, the Caro-Wei sum $W$ bounds the independence number from below, while $(\Delta+1)W/2$ bounds it from above. A second interval has the Havel-Hakimi residue as its lower endpoint and the annihilation number as its upper endpoint. This note proves that the annihilation number does not exceed the Caro-Wei upper endpoint. This upper bound is sharp, with equality for regular graphs of positive degree and even order. Since the residue is at least $W$, this upper bound places the residue-annihilation interval inside the Caro-Wei interval. Both endpoints can improve strictly on one graph, although the sharp worst-case endpoint ratio remains unchanged, even for connected graphs. The annihilation-number bound also proves a conjectured lower estimate based on the annihilation number and residue for every connected graph with at least three vertices. Equality in this estimate holds for exactly eight graph types, while the single edge violates it. For maximum degree at least three, however, the estimate never exceeds the residue, even without connectedness.
- [921] arXiv:2608.03465 (replaced) [pdf, html, other]
-
Title: A Spectral Proximal DC Algorithm with Nonmonotone Line Search and ExtrapolationSubjects: Optimization and Control (math.OC); Numerical Analysis (math.NA)
The paper proposes a novel proximal difference-of-convex (DC) algorithmic framework to solve general nonconvex, nonsmooth optimization problems. The basic pBBDCA method combines Barzilai-Borwein (BB) parameter selection with a nonmonotone directional line search, requiring only one proximal solve per iteration. We also develop two extrapolated variants using restart and safeguarding mechanisms. The global convergence of the proposed framework is rigorously established under the Kurdyka-Łojasiewicz property. Numerical experiments on the SCAD-regularized least squares problem and graph Ginzburg-Landau image segmentation models demonstrate that the proposed methods achieve highly competitive efficiency and accuracy.
- [922] arXiv:2608.14013 (replaced) [pdf, html, other]
-
Title: Kusner's conjecture on equilateral sets is falseComments: The counterexample (Theorem 1) has been formalised in Lean from standard axioms. Lean proof: this https URLSubjects: Metric Geometry (math.MG); Combinatorics (math.CO); Numerical Analysis (math.NA)
We disprove Kusner's 1983 conjecture that every equilateral set in $\ell_p^n$ with $2<p<\infty$ has at most $n+1$ points: there exist $58$ points in $\mathbb{R}^{56}$ whose pairwise $\ell_5$ distances are all equal, so the maximum equilateral-set size satisfies $e(\ell_5^{56})\ge58>57$. This is the first equilateral set of more than $n+1$ points in $\ell_p^n$ for any finite $p\ge2$. The construction persists on an open interval of exponents around $5$. The configuration is specified by the unique solution of an explicit polynomial system with rational coefficients in a rational box; existence and uniqueness are proved by exact integer inequalities. The existence of the $58$-point equilateral set in $\ell_5^{56}$ has been formalised in Lean from standard axioms.
- [923] arXiv:2609.01237 (replaced) [pdf, html, other]
-
Title: The $α$-Limit Problem: Convergence of a Linear Degenerate Interface Transmission ProblemSubjects: Analysis of PDEs (math.AP); Numerical Analysis (math.NA)
We study the singular limit of a family of linear degenerate interface transmission problems arising from a regularization procedure in the newly proposed Two-Parameter Diffuse Domain Method (DDM2p). For $\alpha>0$, the regularized problem admits a strictly convex variational formulation on $H^{1}(\Omega)$. In the limit $\alpha\to0$, the problem degenerates to a weakly coupled interface system with a nonstandard energy structure. To characterize the limit, we introduce a closed Hilbert subspace $\mathcal{H}\subset H^{1}(\Omega)$, defined through an auxiliary Helmholtz problem on an annular subdomain $\Omega_2\subset \Omega$, and identify the limiting energy functional $\mathcal{E}_{0}$ on $\mathcal{H}$. We prove that the regularized energies $\mathcal{E}_{\alpha}$ restricted to $\mathcal{H}$ $\Gamma$-converge to $\mathcal{E}_{0}$ in the strong $L^{2}(\Omega)$ topology, using the standard framework. Consequently, minimizers of $\mathcal{E}_{\alpha}$ converge to the unique minimizer of $\mathcal{E}_{0}$, which is shown to be equivalent to the solution of the limiting interface problem. We further prove a strong convergence $u_{\alpha}\to u_{0}$ in $H^{1}(\Omega)$ and establish a convergence rate of $O(\alpha)$. Numerical experiments in one spatial dimension exhibit first-order convergence, suggesting that the proved $O(\alpha)$ upper bound is sharp.
- [924] arXiv:2609.17726 (replaced) [pdf, html, other]
-
Title: Self-Supervised Learning for Robust Resonance Mass Regression in Cascade DecaysComments: 9th Workshop on Machine Learning and the Physical Sciences (ML4PS 2026)Subjects: High Energy Physics - Experiment (hep-ex); Machine Learning (cs.LG)
Reconstructing the mass of a heavy resonance from its decay products with missing energy is one of the central tasks that directly determine the sensitivity in new physics searches at collider experiments. Supervised learning approaches to this problem often struggle to generalize well due to the presence of various systematic uncertainties and distribution shifts. Exhausting all possible variations in the labeled data can be very compute-intensive, while a failure of the model to generalize can corrupt the reconstructed resonance widths that are critical in peak-hunting analyses. In this work, following the foundation model paradigm, we use a self-supervised approach to pre-train a transformer encoder with VICReg to learn an embedding invariant to various corruptions, then fine-tune it for mass regression on a heavy resonance with masses ranging from 2.5 to 6.5 TeV and a SUSY-like cascade decay into an eleven-body final state. We show that the pre-trained model reconstructs sharper resonance peaks and has a more stable performance under various realistic corruptions, compared to a supervised model of the same architecture trained on the same augmented data from scratch.
- [925] arXiv:2609.17738 (replaced) [pdf, html, other]
-
Title: Similarity Pairing with Energy Mover's Distance for Self-Supervised Pre-Training at the LHCComments: 9th Workshop on Machine Learning and the Physical Sciences (ML4PS 2026)Subjects: High Energy Physics - Experiment (hep-ex); Machine Learning (cs.LG)
Many self-supervised methods for training foundation models at the Large Hadron Collider (LHC) rely on data augmentations to encourage the model to embed events into a representation space invariant to certain physical or detector symmetries. A common challenge arises from the large freedom in choosing a proper set of augmentations on which downstream performance depends. The implementation of augmentations involves either modifying existing events, potentially breaking the event fidelity, or simulating more event variants, which is computationally intensive. In this work, we present a data-driven method of pairing events by their similarity via the energy mover's distance (EMD), which measures how similar two events are in terms of the work required to transform one into the other. With this approach, distinct events are sampled and matched by their similarity to serve as views for learning invariance, keeping the physics content of each event intact without handcrafted distortions. We demonstrate this augmentation-free pairing method by pre-training on QCD jets via self-distillation and show that it can yield semantic jet embeddings with downstream discrimination power comparable to or better than an augmentation-based baseline.
- [926] arXiv:2609.18147 (replaced) [pdf, html, other]
-
Title: Dynamic Pooling and Regional Participation in Deceased-Donor Organ AllocationComments: 47 pages, 3 figuresSubjects: Theoretical Economics (econ.TH); Computer Science and Game Theory (cs.GT)
Moving from geographically fragmented to pooled waiting lists in deceased-donor organ transplantation can improve efficiency, but it raises concerns about regional fairness and participation incentives. This paper studies Pareto gains from such transitions in a multiclass queueing model with impatient agents, perishable items, and a tractable homogeneous compatibility friction. In contrast to standard approaches that emphasize static match quality or overlook regional incentives in dynamic contexts, this analysis identifies dynamic risk pooling as the primary source of Pareto improvements for regions. Pooling reduces organ waste resulting from stochastic supply-demand fluctuations, temporal mismatches, and random compatibility outcomes. Although this dynamic surplus theoretically enables Pareto improvements, a purely utilitarian policy that treats all patients equally across regions often fails to achieve these improvements when regions are asymmetric. In these situations, utilitarian redistribution removes an excessive number of organs from high-supply regions and violates their participation constraints. To address this conflict, it is demonstrated that the set of achievable utilities can be represented by mixtures of static priority index policies. Building on this structure, the Constrained Pareto Frontier Search (CPFS) is proposed as a polynomial-time procedure that computes an approximately optimal allocation policy with arbitrarily small participation-constraint violations, implemented as a single priority list with a lottery at registration. The results indicate that, by adjusting priority indices to ensure redistribution remains within the gains from pooled supply, demand, and compatibility risk, policymakers can achieve global efficiency while maintaining local participation incentives.
- [927] arXiv:2609.38524 (replaced) [pdf, html, other]
-
Title: Generative sequence modeling for infinite memory processes via predictive statesSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME)
We consider estimating the one-step-ahead conditional distribution of a multivariate stochastic process. Many existing approaches rely on assumptions such as finite-range memory, sparsity, or additivity, which can be poorly suited to processes with long-range nonlinear interactions. However, without such structural assumptions, nonparametric estimation is challenging due to the curse of dimensionality. To address this challenge, we introduce a new estimator based on learning representations of the \emph{predictive states} of a stochastic process. We show that the statistical complexity of the estimation problem is determined by the dimension of this predictive state representation. Furthermore, we show that a $p$-dimensional predictive state space can be represented in a $2p+1$-dimensional feature space. Our results imply that faster convergence rates are achievable when the predictive state space is lower-dimensional, i.e., when the past history can be compressed into a lower-dimensional statistic that is sufficient for predicting the future. We establish theoretical convergence rate guarantees for an instantiation of our method based on deep neural network estimators. We implement our method in PyTorch and evaluate its performance on synthetic data and real \emph{C. elegans} data.
- [928] arXiv:2610.07740 (replaced) [pdf, html, other]
-
Title: High-Dimensional Online Calibration from Harmonic WeightsComments: In-depth talk about this work: this https URLSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
We study the online calibration of multidimensional forecasts over an arbitrary convex set $Y\subseteq\mathbb{R}^d$ relative to an arbitrary error norm $\|\cdot\|_L$. For forecasting $d$ binary outcomes simultaneously ($Y=[0,1]^d$), we give the first algorithm that achieves $\varepsilon$-calibration in a number of rounds that is polynomial in $d$ for every fixed accuracy. It requires $d^{O(1/\varepsilon)}$ rounds, exponentially improving the dimension dependence of previous bounds. For multi-class forecasting ($Y=\Delta_d$), we obtain the same $d^{O(1/\varepsilon)}$ rate, improving the $d^{\widetilde{O}(1/\varepsilon^2)}$ bounds of Peng and Fishelson, Golowich, Mohri, and Schneider.
Our algorithm is simple: on each round, it outputs a harmonically weighted distribution over harmonically smoothed past outcomes. The same algorithm works for every forecast set and error norm. More generally, it achieves $\varepsilon$-calibration after $\exp(O(\gamma(Y,L)/\varepsilon))$ rounds, where $\gamma(Y,L)$ is a geometric parameter defined by a matrix discrepancy problem. The harmonic weights are motivated by the fact that the discrete Hilbert transform matrix achieves the optimal discrepancy up to a universal constant, simultaneously for every $L$. This optimality result may be of independent interest.
The same harmonic responses also yield an algorithm for high-dimensional Blackwell approachability. Its guarantees are governed by a sharpened geometric parameter that captures the joint geometry of the outcome and action sets. - [929] arXiv:2610.13357 (replaced) [pdf, html, other]
-
Title: ECGGuard: Deep Watermarking of Electrocardiogram Signals With Error-CorrectingComments: 12 pages, 4 figures, 5 tablesSubjects: Signal Processing (eess.SP); Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV)
Wearable electrocardiogram (ECG) recordings pass through multiple networks and systems before automated interpretation, leaving patient metadata vulnerable to loss. Digital watermarking can embed this metadata in the signal, but conventional methods fail under time shifts, and high waveform similarity does not guarantee the same downstream diagnosis. We present \rev{ECGGuard}, a deep watermarking framework for single-lead ECG. A fully convolutional generator embeds the payload without signal quantization, a detector decodes the message frame by frame and checks for watermark presence, and during training a frozen arrhythmia classifier penalizes changes that would alter its predictions. Because high bit accuracy does not ensure whole-message recovery, the payload is protected by extended Hamming codes, \rev{a validation-selected channel ordering}, and soft-decision decoding. On the MIT-BIH Arrhythmia Database, \rev{averaged over three training runs, ECGGuard} keeps bit accuracy above \rev{98\%} under time shifts of 10 to 50 samples, where transform-domain baselines drop to chance. The coded 16-bit payload is recovered exactly in \rev{99.95\%} of clean segments\rev{, in over 96\%} under time shifts, 10\% sample deletion, or 10\% packet loss\rev{, and in 99.1\% for unseen subjects}. \rev{For the guiding classifier, the diagnostic loss limits the drop in ventricular-beat sensitivity to 1.6 percentage points on the test records (35.0 without it) and to 12.2 points for unseen subjects (67.0 without it); the drop grows to 10.3 points for independently trained classifiers, and the protection does not transfer to an external database.} \rev{ECGGuard} has 0.56 million parameters and needs about \rev{19}~ms per segment on a desktop CPU.
- [930] arXiv:2610.29771 (replaced) [pdf, html, other]
-
Title: Logics vs doctrines, a mirror symmetryComments: Corrected minor mistakes and introduced the notion of tightness in Section 5Subjects: Category Theory (math.CT); Logic in Computer Science (cs.LO); Logic (math.LO)
We study the relationship between doctrines, encoding 2-categories of theories in predicate logic, and fragments of geometric logic, defined through Kan injectivity in the 2-category of topoi. After a brief historical account of the notion of doctrine, we propose our working definition. Our main result then establishes a Galois connection between doctrines and subgeometric logics. This connection suggests a mirror symmetry between the world of doctrines and the world of logics, in which constructions and results on either side admit counterparts on the other. We develop this symmetry through constructions of syntactic categories and classifying topoi in both settings, recovering the familiar examples from categorical logic. We prove paired Diaconescu-type theorems and give symmetric descriptions of the two closure operators induced by the Galois connection. Finally, we characterise when passing from a doctrine to its closure induces a reflection between the corresponding 2-categories of algebras, and we relate this reflection to conceptual completeness.
- [931] arXiv:2610.35581 (replaced) [pdf, html, other]
-
Title: Integrate the Reward, Don't Sample It: Window Likelihood for Interruptions and Backchannels in Full-Duplex Speech ModelsGuoqiang Hu, Lichao Zhang, Yizhou Peng, Yuhao Luo, Haifeng Hu, Zhongqian Sun, Wei Yang, Eng Siong ChngSubjects: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Sound (cs.SD)
When a user talks over a full-duplex dialogue model, it must decide whether the user is interrupting (stop) or backchannelling (carry on). Post-training can improve this discrimination or change how readily it stops. In our GRPO runs the second dominates: no configuration lowers both error rates on more than one model, and two released checkpoints discriminate no better than their bases. We trace this partly to how the reward is estimated. The probability of stopping inside a short window is a known function of a few stop probabilities the policy emits, yet GRPO estimates it by sampling. A group of rollouts that all agree gives no gradient, even when all fail; the push on the stop threshold is strongest where the model is unsure, whichever outcome is right; and whole-rollout credit reaches decisions the reward never judges. Confining the credit to the window removes only that leakage. Window likelihood (WL) samples the words but integrates over the stop decisions, so the probability of the correct outcome is computed exactly and its logarithm maximised. On HumDial-FDBench, across four open models, WL gives the best F1 and lowers both error rates by more than their confidence intervals, without talking into silence.
- [932] arXiv:2610.41180 (replaced) [pdf, html, other]
-
Title: Quantum query complexity of Hurwitz stability testingComments: 36 pages, including appendices; 1 figure, 1 tableSubjects: Quantum Physics (quant-ph); Computational Complexity (cs.CC)
Determining whether every trajectory of the linear system $\dot{x}=Ax$ converges to zero (or, equivalently, whether $A$ is Hurwitz stable) is a fundamental problem in control theory. It arises, for example, in verifying the stability of feedback controllers and state observers, as well as of linearized dynamical models. We study the quantum query complexity of deciding whether a matrix $A\in\mathbb{C}^{n\times n}$ is Hurwitz stable, given an exact normalized block-encoding of $A$. The query complexity is characterized in terms of the system dimension $n$ and a supplied lower bound $0<\nu\leq \omega_A$ on the imaginary-axis distance $\omega_A:=\min_{t\in\mathbb{R}}\sigma_{\min}(A-\mathrm{i}tI)$, where $\sigma_{\min}(A)$ is the smallest singular value of $A$. We prove a query-complexity lower bound of $\Omega(\sqrt{n}/\nu)$, which holds even when $A$ is restricted to be real and diagonal. For general matrices with nonzero $\omega_A$, we encode an approximation of the right-half-plane spectral projector $\operatorname{sign}^+(A):=(I+\operatorname{sign}(A))/2$ as a block of a matrix inverse, where $A$ is Hurwitz if and only if the projector is zero. To test this, we produce a quantum algorithm which determines if a $(\beta,b,1/(4\sqrt{n}))$-block-encoded projector matrix (like $\operatorname{sign}^+(A)$) is zero or not; for constant success probability, it has complexity $\Theta(\beta\sqrt{n})$. When the zero test is applied to specific approximations of $\operatorname{sign}^+(A)$, this yields an upper bound of $O(\sqrt{n}\nu^{-2}\log(\sqrt{n}/\nu))$ queries in the general case, and $O(\sqrt{n}\nu^{-1}\log n)$ when $A$ is Hermitian or normal. The latter upper bound matches the lower bound up to a logarithmic factor. Thus the dimension dependence of quantum Hurwitz testing is nearly optimal, while the remaining factor in the non-normal case isolates an open conditioning gap.
- [933] arXiv:2611.11985 (replaced) [pdf, html, other]
-
Title: Trajectory Density Ratios under Weak Support: A Survey across the SciencesSubjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Many problems across the sciences compare two stochastic processes through the sequences they generate. Reinforcement learning asks what a new policy would return from the trajectories of an old one; molecular simulation asks how much more a ligand binds from samples of the unbound state; climate attribution asks how a heat wave shifted under warming; and language-model post-training asks what a deployed agent will do from logged interactions. In every case, one quantity would answer all these questions at once: the ratio of the two probability laws over their generated sequences, the trajectory density ratio. This ratio exists only under strong conditions. When one process reaches states the other cannot, the ratio does not exist; when their overlap is thin, its variance grows without control with the sequence length; and when they concentrate on different lower-dimensional sets, the ratio is undefined altogether. Weak support of this kind is the norm for sequence data, not a sampling accident, and its price is paid in identifiability rather than in noise alone. This survey organizes the field around a three-level diagnosis of weak support and a map of the nine retreats that methods take when the full ratio is out of reach, from direct estimation and reweighting through staged decomposition, stationarization, and partial-identification bounds to smoothing, transport couplings, operator spectra, and direct estimation of a single functional. Each retreat keeps part of what the ratio offers and pays for the rest, and each application domain has standardized the retreat its data force. The survey closes with open problems: testable diagnostics at every level, partial identification at the trajectory level, a rate theory of time-amplified error, and benchmarks with ground truth.
- [934] arXiv:2611.13971 (replaced) [pdf, html, other]
-
Title: TransGHOST: Real-Time, Synthesis-Free Reconstruction of Instrument-Occluded Regions in Ophthalmic MicrosurgeryComments: 20 pages, 10 figures, 11 tablesSubjects: Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
Intraoperative scene perception in ophthalmic microsurgery is limited by instrument occlusion: when a tool enters the high-magnification field, it conceals tissue from both the surgeon and the perception components of robot-assisted systems that process the video. Generative inpainting can remove the occluder but replaces the missing image content with synthesis, which is unsuitable when visual provenance matters. Many video inpainters also depend on future frames and cannot operate online. We present TransGHOST, a causal reference-guided reconstruction pipeline whose non-generative safe mode operates in real time. Every restored value is obtained by resampling a recorded frame using geometric transforms fitted only to visible regions. The system provides per-pixel confidence and provenance and marks unsupported pixels as abstentions. We evaluate TransGHOST on a new veterinary clinical corpus of canine cataract surgery with pixel-level instrument and anatomy labels. Only the occluded region is scored, and each scoring target is verified to differ from the reference supplied to the system. Safe mode reaches 31.89 dB masked PSNR and 0.915 masked SSIM at 51.7 ms median and 85.5 ms 95th-percentile latency. On a pre-declared held-out cohort using instrument cutouts excluded from segmentation training, it reaches 29.83 dB masked PSNR, while matched-mask evaluation on validation shows a 3.45 dB advantage over LaMa. Region-aware reference selection improves masked PSNR by 3.60 dB over recency when candidate references can themselves be occluded. Reference-age analysis further defines the operating interval for the brief, repeated occlusions targeted by TransGHOST.
- [935] arXiv:2611.14182 (replaced) [pdf, html, other]
-
Title: Quantum-Assisted Counterfactual Regret Minimization for Bounded-Payoff Imperfect-Information GamesSubjects: Quantum Physics (quant-ph); Emerging Technologies (cs.ET); Computer Science and Game Theory (cs.GT)
Imperfect-information games with a graded-reward structure present significant computational challenges for equilibrium computation, since the payoff expectations must be estimated across exponentially large strategy spaces. Counterfactual Regret Minimization over the years has proven effective for approximating Nash equilibria, but Monte Carlo variants create a large overhead of sampling costs when the counterfactual values are estimated from sampled game-tree trajectories. We propose a novel Quantum-Assisted Counterfactual Regret Minimization, a framework that uses Quantum Amplitude Estimation to reduce the computational complexity of the regret updates in bounded-payoff imperfect-information games, while preserving the convergence of its classical counterpart. We give a theoretical analysis of the resulting exploitability bound, discuss NISQ-compatible amplitude-estimation variants, the structure of their resource requirements, and outline what is needed to validate the potential advantage empirically.
- [936] arXiv:2611.17067 (replaced) [pdf, html, other]
-
Title: An Invariant Neural-Network Eddy-Viscosity Closure with Iterative Solver Coupling for Molten Salt Fast Reactor Thermal-HydraulicsZaid Abulawi (1), William Oliveira (2), Akzhol Almukhametov (1), Jun Fang (3), Ling Zou (3), Rui Hu (3), Stefano Lorenzi (2), Yang Liu (1) ((1) Department of Nuclear Engineering, Texas A&M University, (2) Department of Energy, Nuclear Engineering Division, Politecnico di Milano, (3) Nuclear Science and Engineering Division, Argonne National Laboratory)Comments: 42 pages, 12 figuresSubjects: Fluid Dynamics (physics.flu-dyn); Machine Learning (cs.LG)
System thermal-hydraulics codes such as the System Analysis Module (SAM) close turbulence on coarse meshes with algebraic zero-equation mixing-length models. These models cannot represent the non-equilibrium turbulence, flow separation, and large-scale recirculation of open-cavity advanced reactors. We develop an invariant neural-network eddy-viscosity closure for SAM and apply it to the Molten Salt Fast Reactor (MSFR). The closure predicts the turbulent kinetic energy and timescale of each element from the flow and geometry of its neighbors, without solving turbulence transport equations. The resulting non-negative eddy viscosity enters the implicit diffusive operator, which keeps the momentum equations well conditioned. We prove that the closure is exactly invariant to translation, Galilean boost, rotation, reflection, and neighbor ordering. A network trained on high-order Nek5000 fields sees a different feature distribution on the coarse SAM mesh. We remove this covariate shift by iteratively retraining the network on SAM flow states against fixed Nek5000 targets. For a held-out steady state, the coupled closure reproduces the single recirculation cell and eddy-viscosity plume of the Nek5000 reference. In a 120 s pump-coastdown transient, it halves the velocity error of the mixing-length model, from 37.6% to 17.7%, and matches the Nek5000 eddy viscosity with a time-averaged relative error of 12%. The remaining velocity error comes from the coarse mesh. The coupled transient takes about 0.2 core-hours, compared with about 330 core-hours for Nek5000. The results show that the closure can recover two-equation RANS turbulence behavior in a coarse-mesh system code at a small fraction of the CFD cost.