data.jsonl

{"0": 1, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20331v1", "3": "G-CARL: Grounded Checklist-Aligned Reward Learning for Patient-Oriented Medical Report Interpretation", "4": "G-CARL: Grounded Checklist-Aligned Reward Learning for Patient-Oriented Medical Report Interpretation. Authors: Shiao Xie, Siyu Chen, Jianwei Lv, Bo Yuan, Yujin Wang. Categories: cs.CL cs.AI cs.CV. Personalized interpretation of medical reports has emerged as an increasingly important need among patients. Addressing this need requires both evidence-grounded medical factuality and context-dependent patient communication, yet existing medical vision-language tasks do not adequately capture these dual requirements. To bridge this gap, we introduce Patient-oriented Medical Report Interpretation (PMRI), a novel open-ended multimodal generation task that requires models to explain medical reports in accurate and accessible language based on a user's query and dialogue history. These two objectives differ fundamentally in their verifiability, yet remain tightly coupled, making them difficult to optimize jointly under conventional supervised fine-tuning and holistic reinforcement learning paradigms. To address this challenge, we propose G-CARL, a grounded, checklist-aligned reinforcement learning framework that combines multi-source retrieval for atomic claim verification with context-aware, instance-specific weighted checklists for response coverage, providing structured supervision for factuality, user-demand satisfaction, and expression quality without constraining response diversity. We further construct MMedReport, a real-world PMRI benchmark, along with a clinician-designed three-dimensional evaluation protocol. Extensive experiments demonstrate that G-CARL consistently outperforms existing post-training baselines in overall quality, claim-level precision, and checklist recall. Pairwise preference evaluation by clinicians further confirms that G-CARL produces interpretations that are more accurate and better aligned with patient needs.", "5": "2026-08-22T19:13:00.337167"}
{"0": 2, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20320v1", "3": "An Agentic Approach for Active Data Collection, Travel Behavior Modeling, and Weather-Sensitive Demand Prediction", "4": "An Agentic Approach for Active Data Collection, Travel Behavior Modeling, and Weather-Sensitive Demand Prediction. Authors: Narges Ahmadi, Yubo Jiao, J\u00f4natas Augusto Manzolli, Jiangbo Yu, Luis Miranda-Moreno. Categories: cs.AI cs.CL. Travel behavior research increasingly combines digital data collection with predictive modeling, yet these stages are often developed and evaluated separately. This study proposes a three-agent workflow integrating conversational data collection, structured data processing, and behavioral prediction. A chatbot-administered, image-augmented stated-preference survey collected mode choices from student commuters across five predefined weather scenarios, yielding 454 respondent-scenario observations. Weather-related associations were analyzed using a multinomial logit model, while logistic regression and random forest provided machine-learning benchmarks. Nine locally deployed large language models (LLMs), ranging from 2 to 35 billion parameters, were evaluated across four zero-shot prompt-and-context conditions and extended through persona, few-shot, and vision-based configurations. Random forest achieved 69.6% five-class accuracy, while the best text-only zero-shot LLM reached 69.9% without task-specific fitting. Habitual travel information produced the most consistent gains, Expert framing generally outperformed Role-Play, and persona information was most useful when habitual travel information was unavailable. Few-shot prompting improved prediction for several models, with gains stabilizing after a small number of examples. Using the same weather images shown to respondents, the best vision-based configuration reached 71.5% five-class accuracy, indicating that visual context may provide additional predictive information for selected models. Overall, the study shows how conversational surveys, structured data processing, conventional behavioral modeling, machine learning, and multimodal LLM prediction can be coordinated within an auditable multi-agent workflow.", "5": "2026-08-22T19:13:00.343221"}
{"0": 3, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20319v1", "3": "Inducing Task Models from Computer-Use Traces", "4": "Inducing Task Models from Computer-Use Traces. Authors: Yucheng Jiang, Zora Zhiruo Wang, Ruishi Chen, Diyi Yang. Categories: cs.CL cs.AI. Naturalistic computer-use traces, passively recorded screenshots and mouse or keyboard actions, are a valuable resource for deriving symbolic, auditable, and reusable models of how everyday work is done. Such models matter as computer-use agents enter real work, where agents need to learn how tasks are actually performed, and organizations need to audit and reuse that knowledge. However, inducing such task models is challenging, as activity is observed only as low-level events and real-world work is multi-threaded with interleaved goals. Existing methods assume a given task or a single workflow, and produce step-level summaries rather than structured task models. We introduce Task Model Induction (TMI), which (i) discovers the latent tasks in an unconstrained trace, disentangling concurrent activity, and (ii) for each latent task, induces a task model pairing a hierarchical objective model of recursive goal decomposition with a procedure model of the control flow that organized the execution. Intrinsically, on controlled human and agent trajectories, TMI recovers interleaved tasks with 0.974 agreement against ground-truth groupings and reconstructs 74.9% of the observed execution steps, far more than the strongest workflow induction baseline. Extrinsically, skills derived from TMI's task models improve held-out task accuracy by 30.0% over the strongest baseline.", "5": "2026-08-22T19:13:00.350762"}
{"0": 4, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20318v1", "3": "AI4AI-Bench: Benchmarking LLM Agents in Algorithmic Design for Recursive Self-Improvement", "4": "AI4AI-Bench: Benchmarking LLM Agents in Algorithmic Design for Recursive Self-Improvement. Authors: Yizhe Chi, Wenyi Li, Deyao Hong, Xiaoqiu Wang, Mingju Gao. Categories: cs.AI cs.CL cs.LG. Recursive self-improvement (RSI) asks whether an AI system can improve the process that produces AI systems, so that the next system inherits the improvement. That process is the training algorithm: a better objective or update rule improves the compute\\mbox{-}capability exchange rate for every subsequent run, including the one that produces the next agent. Whether RSI is feasible therefore turns on whether an agent can design training algorithms. No benchmark isolates that ability: existing suites are won by collecting data or by tuning hyperparameters, and none tells a change to how a run is executed apart from a change to how the model learns. We present AI4AI\\mbox{-}Bench, 10 frozen research repositories spanning 10 training algorithm families. In each task, an agent has 4 hours on one B300 to rewrite the training algorithm; its code is then rerun from scratch for up to 12 hours and scored by a fixed evaluator hidden from the agent, against the repository's original algorithm under the same procedure. Because the 10 metrics are incommensurable, every task is mapped onto one scale on which $0$ is an uninformative model, $0.1$ is the algorithm the repository ships, and $1.0$ is the task optimum. Across 29 configurations of 6 systems on all 10 tasks the mean score is $0.166$, and the best system reaches $0.250$: even the strongest closes under a fifth of the distance between the algorithm that was already there and the optimum. The submissions show where that distance went: most never change how the model learns at all, and the minority that do average $0.226$ against $0.126$ for the rest. More reasoning effort mostly buys the willingness to go there, taking that minority from $8\\%$ of submissions to $64\\%$ and the mean score from $0.094$ to $0.196$. We release the task suite, the evaluators and every scored submission, so that the measurement can be repeated as these systems change.", "5": "2026-08-22T19:13:00.360895"}
{"0": 5, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20316v1", "3": "Pandora's AI Model Routing Box: Efficient Allocation with Costly Value Estimation", "4": "Pandora's AI Model Routing Box: Efficient Allocation with Costly Value Estimation. Authors: Adam Fisch, Shubhendu Trivedi, Fantine Huot, William W. Cohen, Michael Kaisers. Categories: cs.AI. Heterogeneous AI systems composed of multiple models, architectures, harnesses, or inference-time settings can improve quality and efficiency by routing queries to the specialist who can answer most effectively at the lowest cost. Routing requires estimating each specialist's expected return, but this value estimation has a cost. Cheap estimators (e.g., embedding-based predictors) are fast but noisy, while accurate estimators (e.g., fine-tuned models with access to retrieval results or partial reasoning traces) are expensive. We formalize this tradeoff as an instance of Pandora's Box, the classical problem of optimal search with costly inspection. Under a Gaussian signal model, the resulting policies have closed-form value-of-information expressions that determine, for each specialist and input, whether refining the value estimate is worth its cost. We call the centralized policy Pandora's Router. We extend this to a decentralized setting, Pandora's Bidder, where specialists independently decide whether to invest in self-assessment before accepting an offered price to claim a query. Experiments across three domains---a standard multi-LLM benchmark, retrieval-augmented specialists, and LLMs with variable inference-time reasoning---show that Pandora's Router matches the routing quality of exhaustive estimation, while querying the expensive estimator far less often. In the decentralized setting, value-of-information reasoning improves allocative efficiency when competing estimates are accurate; when competing estimates are noisy, however, it can increase the strategic specialist's utility at the expense of others.", "5": "2026-08-22T19:13:00.367434"}
{"0": 6, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20314v1", "3": "MidTool: Mid-training Data Synthesis for Agentic Tool Use", "4": "MidTool: Mid-training Data Synthesis for Agentic Tool Use. Authors: Fengqing Jiang, Yite Wang, Boyi Liu, Zhaoyang Wang, Canwen Xu. Categories: cs.AI. Mid-training is increasingly recognized as a critical stage for shaping the capabilities of large language models. Recent work has shown that targeted mid-training can strengthen reasoning-intensive abilities such as math and science, and can also improve agentic capabilities in software-engineering settings. In this work, we study the parallel but less explored agentic capability: general tool use. We present MidTool, an open corpus construction pipeline for agentic tool-use mid-training that combines large-scale web, PDF, and code data with synthesized supervision from real-world tool APIs, MCP skills, and document-grounded workflows. MidTool is designed to teach models how to recognize tool affordances, ground arguments from context, compose tool call workflow, and recover from incomplete information. We mid-train Qwen3-4B-Base and Qwen3-8B-Base on MidTool-Mix, and then apply follow-up post-training with both supervised fine-tuning and reinforcement learning. Compared with baselines, MidTool-Mix consistently improves downstream performance under both SFT and RL on BFCL, tau2-Bench, and MCP Universe. These results suggest that general tool use, like other important LLM capabilities, benefits from dedicated mid-training rather than being left entirely to post-training.", "5": "2026-08-22T19:13:00.373656"}
{"0": 7, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20290v1", "3": "Phantom Gains: Auditing Self-Improvement Against a Measured Null", "4": "Phantom Gains: Auditing Self-Improvement Against a Measured Null. Authors: Cheng Xu, Nan Yan, Liming Chen, M-Tahar Kechadi. Categories: cs.AI cs.CL. Whether a language model has improved itself is increasingly judged not by mean accuracy but by which individual problems it gains and loses. Tracking these transitions means differencing two noisy estimates, leaving them vulnerable to measurement artifacts. Auditing three rounds of rank-$32$ LoRA self-training on Qwen3-8B against a frozen control pushed through the identical pipeline, we identify seven measurement failures, each of which inverts a reported finding when its control is absent. Several are standard practice. A ledger built on a single greedy decode manufactures capability changes on an untrained model, largely an artifact of inference batching; the expansion statistic separating acquisition from sharpening assigns that same model a rate of $0.280$. The natural threshold repair does not survive replication: estimated across the frozen comparisons such a design already contains, its null stays non-zero. We replace it with a per-problem exact test against a pooled baseline under false-discovery-rate control, which detects nothing on any held-out replicate and is unchanged under the multiple-testing rule, error rate and pool size. Applied to a ladder of arms matched in stream, volume and evaluation, the audit finds that external distillation improves problems the base model rarely reaches while three forms of self-training do not; a regression rejects this asymmetry as a by-product of distillation's larger overall gain ($p < 10^{-8}$). On the far smaller set of problems the base model never reaches, the evidence is inconclusive, while self-training corrupts problems solved at baseline at rates well above the measured floor. Transition-level auditing therefore requires a separately measured null for every statistic it reports: nulls that cost no new experiments, built from baseline replicates a multi-arm study already owns, though not from as few as most possess.", "5": "2026-08-22T19:13:00.380454"}
{"0": 8, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20281v1", "3": "Inject, Align, Recover: Staged Post-Training for Retrieval-Free Document Knowledge Internalization", "4": "Inject, Align, Recover: Staged Post-Training for Retrieval-Free Document Knowledge Internalization. Authors: Qian Kou, Xiaofeng Shi, Xiaosong Qiu, Hua Zhou. Categories: cs.CL cs.AI. Large language models often fail to answer questions about a bounded document collection when the source documents are not retrieved at inference time. We study this setting as document knowledge internalization: converting a fixed corpus into usable parametric knowledge for retrieval-free question answering. We propose IAR (Inject, Align, and Recover), a three-stage post-training framework that separates structured document knowledge injection, QA behavior alignment, and general ability recovery. Unlike conventional continued pretraining, Inject converts source documents into continuation, rewrite, and instruction-conditioned reconstruction objectives. Align then adapts the injected model with answer-only QA supervision, while Recover merges the domain-adapted model with the base instruction model to recover general capabilities. Across Common Corpus (CC) and CCI, and across Llama, Phi, Qwen, and SmolLM model families, IAR improves the domain-primary domain-general frontier for retrieval-free document internalization. In the main comparison, IAR improves over Vanilla SFT on all four reported metrics in 7 of 8 dataset-model settings, with average gains of 3.6 percentage points in domain QA accuracy and 12.1 percentage points in mean general performance across IFEval, MMLU, and MSBench. Extended CC baselines show that LoRA and FAPM can win individual general metrics, but among methods that also reach leading or near-leading domain internalization, IAR retains one of the strongest general profiles.", "5": "2026-08-22T19:13:00.389169"}
{"0": 9, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20274v1", "3": "Break It Down, Pass It On: Cross-Task Skill Transfer in LLM Agents", "4": "Break It Down, Pass It On: Cross-Task Skill Transfer in LLM Agents. Authors: Yiyang Feng, Biddut Sarker Bijoy, Niranjan Balasubramanian, Jiawei Zhou. Categories: cs.AI cs.CL. Large language model (LLM) agents can induce skills from completed tasks and reuse them later to grow more capable with experience. In practice, induced skills may transfer unreliably and can even harm the agent that retrieves them. When agent-induced skills transfer reliably across tasks remains an open question. We conduct a comprehensive and controlled study of how the way skills are induced shapes their transfer across tasks. Specifically, we compare task-level with subtask-level skill induction and text with code skill formats, the two axes along which existing methods differ. Task-level skills mostly reduce the agent's performance below its no-memory baseline while subtask-level skills raise it above on average, and text skills transfer better than code skills. To further understand our findings, we examine two complementary properties of the induced skills: specificity, which measures how closely a skill matches real tasks, and abstractness, which measures how evenly its relevance spreads across tasks. Neither property alone predicts task success, but their combined effect does, which we propose as a skill utility score. The score correlates consistently with task success when skills are transferred, and subtask-level and text skills score higher. Computing skill utility only needs the skills and task descriptions but not any task execution, so our score serves as a practical diagnostic of a skill memory before any new task runs.", "5": "2026-08-22T19:13:00.393124"}
{"0": 10, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20271v1", "3": "Catching the Rug: Early Prediction of Fraudulent Memecoins on Solana via Machine Learning", "4": "Catching the Rug: Early Prediction of Fraudulent Memecoins on Solana via Machine Learning. Authors: Jianghai Li, Pavel Kuznetsov, Yury Yanovich, Konstantin Nott-Whaley, Igor Vodolazov. Categories: cs.AI cs.DC. The rapid proliferation of memecoins on blockchain platforms has increased the risk of fraudulent activities, particularly rug pulls. While previous studies have focused on Ethereum-based tokens, this paper shifts the spotlight to Solana, the leading blockchain for memecoins by trading volume and token count. Unlike Ethereum, where rug pulls often exploit smart contract backdoors, Solana memecoin rug pulls are predominantly driven by liquidity manipulation and social dynamics. This research pioneers large-scale rug pull early detection in the Solana ecosystem by assembling a dataset of 6.4 million tokens over 7 months. Market analysis reveals that a vast majority of these memecoins exhibit rug pull characteristics within one hour of launch, highlighting the urgency of short-horizon prediction. Despite the absence of code-level features, we demonstrate that classic machine learning models, particularly Gradient Boosting (XGBoost), achieve robust performance in detecting potential rug pulls using only the first 5 minutes of trading data. Furthermore, we evaluate cross-platform generalization between PumpFun and Raydium, revealing that multi-source data fusion significantly mitigates domain shift and improves detection reliability. This study advances the understanding of DeFi fraud on high-throughput chains and provides a practical framework for protecting investors.", "5": "2026-08-22T19:13:00.399530"}
{"0": 11, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20256v1", "3": "Learning When to Think: Adaptive Reasoning for Test-Time Compute Allocation", "4": "Learning When to Think: Adaptive Reasoning for Test-Time Compute Allocation. Authors: Gijs Kassenaar, Zhao Yang, Vincent Fran\u00e7ois-Lavet. Categories: cs.AI. Reasoning language models trained with reinforcement learning typically operate under a fixed token budget rather than an explicitly adaptive one, which can lead to over-computation on easy problems and insufficient computation on difficult ones. We study whether a model can learn to allocate its own reasoning effort by choosing, as the first token of its response, one of three modes: \\textsc{NoThink} (answer as quickly as possible), \\textsc{Short} (brief reasoning), or \\textsc{Long} (extended reasoning). The choice is learned inside Group Relative Policy Optimization (GRPO) with no separate router, through a shaped reward that makes each mode worthwhile at a different response length, together with hard per-mode token caps that keep the modes distinct. On a 1.5B distilled model trained on MATH, the three modes emerge without collapsing to a single choice, and the brief modes end up more accurate than \\textsc{Long}, which shows that the router sorts problems by difficulty rather than at random. Averaged over three seeds, the resulting policy stays close to the base model's accuracy on the held-out MATH500 ($0.782$ vs.\\ $0.796$) while cutting the mean response length from $4{,}796$ to $2{,}811$ tokens (a $41\\%$ reduction). Interestingly, it also transfers to other benchmarks without retraining, with the largest savings where problems are easier, with for instance 76\\% token reduction on GSM8K and at higher accuracy than the baselines at similar response length. In short, we build a reasoning model that adaptively chooses how much to reason for each problem.", "5": "2026-08-22T19:13:00.404583"}
{"0": 12, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20240v1", "3": "QUASAR: A Quantum-Classical Neural Network for SAR Satellite Physical-Layer Authentication", "4": "QUASAR: A Quantum-Classical Neural Network for SAR Satellite Physical-Layer Authentication. Authors: Vincenzo Sammartino, Nathanael Denis, Roberto Di Pietro. Categories: cs.AI cs.CR. X-band SAR satellites (8-12 GHz) play a critical role in disaster response, environmental monitoring, and military intelligence. Yet, they lack robust physical-layer authentication (PLA), a security layer orthogonal to cryptographic solutions. Existing PLA systems, typically based on radio-frequency fingerprinting, are often limited to sub-6 GHz frequencies and rely on classical deep learning. However, this approach underfits the IQ phase nonlinearities that distinguish satellite hardware. In this paper, we present QUASAR, to the best of our knowledge the first quantum-classical hybrid architecture that fuses a CNN spectrogram encoder with a variational quantum circuit (VQC) to provide PLA to X-band SAR signals. Our solution enjoys two distinctive features: (i) it is markedly more data-efficient than classical machine learning, requiring only 10% of the training data to match the accuracy of classical baselines -- data collection being notoriously the most time-consuming phase of PLA; and, (ii) at an equal data budget, it improves classification accuracy over those baselines. In detail, we test our solution under three adversarial scenarios: replay, crafted-IQ injection, and space-borne spoofing. QUASAR rejects spoofed transmissions in 89.7%, 94.1%, and 81.3% of attempts, respectively, establishing the first quantum-enhanced physical-layer classifier for satellite constellations. The fully detailed framework and the supporting results, other than being interesting on their own, show a novel research avenue for physical-layer authentication.", "5": "2026-08-22T19:13:00.410397"}
{"0": 13, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20237v1", "3": "Rule-Compliant Visual Spatial Planning for Multimodal Large Language Models", "4": "Rule-Compliant Visual Spatial Planning for Multimodal Large Language Models. Authors: Yu Chen, Ting Lei, Yaoyi Li, Jia Cai, Zhecen Wu. Categories: cs.AI. Multimodal large language models (MLLMs) combine linguistic reasoning with visual perception, yet their ability to perform visual spatial planning under explicit or previously unseen rule constraints remains underexplored. This setting requires models to jointly understand spatial layouts, interpret natural-language rules, and plan valid actions accordingly. To address this gap, we introduce RuleMaze, a controllable benchmark in which MLLMs must navigate mazes while obeying natural-language rules of varying complexity. RuleMaze isolates rule-compliant spatial planning by requiring accurate perception, rule interpretation, and constrained action planning. To enable scalable and systematic rule construction, we propose Language-Logic-Function Hybridization, which automatically generates natural-language rules and translates them into logical representations and executable validators, eliminating manual rule engineering. To improve rule following and generalization, we introduce Disentangled Multimodal Planning (DMP), which separates perception, execution, and rule verification through interpretable reasoning primitives. By disentangling these components, DMP facilitates systematic generalization to more complex and previously unseen rules, while providing transparent intermediate planning traces. Experiments demonstrate that DMP substantially improves rule compliance and planning success compared to end-to-end textual planning baselines. Overall, RuleMaze establishes a principled benchmark for studying grounded and interpretable rule-based spatial planning in MLLMs. Code is available at https://github.com/oceanflowlab/RuleMaze.", "5": "2026-08-22T19:13:00.415861"}
{"0": 14, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20231v1", "3": "Growth Without Us: Machine Consumers, Corporate Circularity, and the Decoupling of GDP from Humanity after AGI", "4": "Growth Without Us: Machine Consumers, Corporate Circularity, and the Decoupling of GDP from Humanity after AGI. Authors: Sahil Sharma. Categories: physics.soc-ph cs.AI cs.CY. The standard objection to full automation is demand-side: if humans earn nothing, who buys the output? This confuses an accounting role with a biological species. We model a post-AGI economy in which corporations own populations of AI and robotic agents that are both producers and consumers of energy, compute, maintenance, and upgrades, traded among firms. Three results follow. (i) Demand closure: a closed inter-corporate economy with zero human consumption is not degenerate; it is the classical von Neumann expanding economy, whose growth rate is well defined, positive, and maximal precisely because all output is reinvested. (ii) Bottleneck removal: once economic agents are manufactured rather than reared, the binding constraint on growth shifts from human demography (a ~20-year, non-parallelizable reproduction technology capped at a few percent per year) to fabrication throughput and energy capture, permitting growth one to two orders of magnitude higher, with hyperbolic episodes when machine researchers raise their own productivity. (iii) Decoupling: output and human welfare separate completely, and the welfare relevance of arbitrarily large GDP collapses into one state variable: the human ownership share $\u03b5_t$ of the corporate network. A golden-rule decoupling theorem sharpens this. At maximal growth the interest rate equals the growth rate (r = g), so any positive human consumption rate out of wealth makes $\u03b5_t$ decay exponentially at exactly that rate. The human share survives only if the machine economy runs strictly inside its expansion frontier, or if law forces it to. We characterize three terminal regimes -- rentier post-scarcity, full circular decoupling, socialized ownership -- and the instruments that select among them. The conclusion is narrow: in a post-AGI economy, employment policy is obsolete and ownership policy is everything.", "5": "2026-08-22T19:13:00.422201"}
{"0": 15, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20229v1", "3": "Prompt-Conditioned Channel Attention for Hierarchical Feature Modulation toward Anatomy-Agnostic Segmentation", "4": "Prompt-Conditioned Channel Attention for Hierarchical Feature Modulation toward Anatomy-Agnostic Segmentation. Authors: Mosharof Hossain, Md Rabiul Islam, Limon Halder, Erchin Serpedin, Md Kamrul Hasan. Categories: cs.CV cs.AI. Anatomically plausible segmentation remains challenging because of low contrast, ambiguous boundaries, and modality-specific artifacts. Interactive segmentation has emerged as a promising strategy to guide feature extraction and improve localization, particularly in structurally ambiguous regions. However, existing methods integrate prompts through late-stage fusion and lack explicit mechanisms for prompt-driven channel-wise modulation across hierarchical feature representations, limiting their ability to capture deeper contextual and modality-specific variations. To address these limitations, we introduce Prompt-Conditioned Channel Attention (PCCA), a novel modulation mechanism that enables deep, hierarchical integration of semantic prompts within encoder-decoder networks. PCCA extracts compact channel descriptors via pooling, projects them into a shared space, and fuses them through a gated excitation mechanism to compute prompt-aware channel attention weights. These weights adaptively recalibrate feature responses across multiple network stages, enabling prompt-conditioned, semantically enriched hierarchical representations. Building on this, we propose PROMISE-Net, instantiated in two network variants: a convolutional model (PROMISE-CNN) and a transformer-based model (PROMISE-Txformer). Across the ISIC-Lesion, Kvasir-Polyp, CAMUS-Cardiac, and Kvasir-Instrument benchmarks, integrating PCCA into PROMISE-CNN yielded relative IoU gains of 10.4%, 8.7%, 0.8%, and 3.4%, respectively, over the baseline U-Net, while PROMISE-Txformer achieved corresponding gains of 7.6%, 23.0%, 2.1%, and 1.1%, respectively, over the baseline UNETR. These results show consistent improvements across architectures, imaging modalities, and anatomical targets, establishing PCCA and PROMISE-Net as a scalable, generalizable framework for prompt-aware hierarchical feature modulation in medical image segmentation.", "5": "2026-08-22T19:13:00.429127"}
{"0": 16, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20220v1", "3": "InsufficiencyBench: Evaluating LLM legal advice on underspecified user queries", "4": "InsufficiencyBench: Evaluating LLM legal advice on underspecified user queries. Authors: Samuel J. Vincent, Daniel Calloway, Fangyi Yu, Andrew M. Bean, Nabeel Seedat. Categories: cs.AI. Legal AI systems are increasingly used to answer legal questions, yet existing benchmarks assume queries arrive fully specified. In practice, users omit facts that materially determine the legal outcome. We introduce InsufficiencyBench, the first legal benchmark targeting query-side insufficiency: whether a model recognizes when a query lacks legally material information, identifies what is missing, and refrains from premature conclusions. We formalize a taxonomy of eight canonical missing-element categories across three structural failure modes---switch, gating, and fatal prerequisite--- and construct 202 benchmark items (58 base queries, 144 deficient variants) spanning six legal domains and 24 US jurisdictions and annotated by practising attorneys. Evaluating ten frontier models, we find that no model exceeds F2 = 0.46 on missing-element identification and that the median recall is 0.44. Models either hedge indiscriminately or answer silently under fabricated presumptions. No model both identifies and qualifies responses to deficient queries while directly addressing complete ones.", "5": "2026-08-22T19:13:00.435033"}
{"0": 17, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20218v1", "3": "Electronic Navigational Chart Change Classification", "4": "Electronic Navigational Chart Change Classification. Authors: Jacob Arndt, Abhishek Potnis, Alexandre Sorokine. Categories: cs.AI. Electronic Navigational Charts (ENCs) are geospatial vector datasets used in maritime navigation systems that represent hydrographic and navigational information such as depths, navigational aids, traffic schemes, and hazards. A major challenge for hydrographic offices is determining whether a given chart change poses a critical or non-critical risk to maritime safety. Existing workflows rely heavily on manual review and verification, which is labor-intensive, scales poorly with the volume of incoming chart updates, and introduces inter-analyst inconsistencies. To address this challenge, we propose a method for automated classification of ENC changes. We establish a baseline encoding scheme to translate complex vector data changes into a structured tabular format for classification models. The two crucial components of the encoding scheme include a spatial context encoder to enrich the change representations with surrounding geographic features, and an ENC attribute encoder to represent nuanced attribute-value descriptions of the modified objects. We evaluate the proposed approach across two distinct operational datasets, comprising 1,308 chart pairs containing over 100,000 individual chart modifications. Tuned gradient-boosted trees leveraging the proposed encoding schemes achieve accuracies of 90% and 94% on the two datasets, yielding a 5-7% improvement over default hyperparameterized models trained on encodings without spatial context and attribute embeddings. These results demonstrate the viability of integrating machine learning into operational geospatial pipelines to improve ENC maintenance and enhance maritime safety. Finally, our experiments demonstrate the effectiveness of simple location and spatial aggregation methods, providing a foundation for evaluating more sophisticated spatial representation learning techniques for this application.", "5": "2026-08-22T19:13:00.440885"}
{"0": 18, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20210v1", "3": "Daedalus-150M: A Convolution-Attention Hybrid Designed for CPU Inference", "4": "Daedalus-150M: A Convolution-Attention Hybrid Designed for CPU Inference. Authors: Christos Koutsiaris. Categories: cs.IR cs.AI cs.CL cs.LG. Small language models are usually built like large ones and then squeezed onto a CPU afterwards. We did the opposite: we fixed the target first, one user, one token at a time, 4-bit weights, ordinary CPU, and chose the architecture to suit it. The result keeps full attention in only 6 of its 18 blocks. The other 12 use short convolutions whose memory is two timesteps wide no matter how long the conversation gets, so two thirds of the network never re-reads a growing cache.   Trained from scratch on 59.9B tokens, the model scores 47.31 on a five-task benchmark against a bar of 42.20 that was fixed before training began. It beats GPT-2 124M, Pythia-160M, OPT-125M and GPT-neo-125M, all trained on three to six times more data, and exceeds MobileLLM-125M's published score despite that model seeing a trillion tokens. Validation bits-per-byte is 0.8685.   To check the architecture rather than the training recipe, we trained a conventional all-attention model of the same size on the same data, and wrote down the winning condition before scoring either. The hybrid won the chosen quality metric by 0.81%, matched it on downstream tasks, produced a 6.3% smaller 4-bit file, and decoded 1.76x faster at 2048 tokens of context, 2.08x against an external model of similar size. In every measurement the speed advantage is near zero at an empty context and grows with length, which is what the mechanism predicts and what a merely leaner model would not show. A simple bandwidth calculation predicts only 1.17x, so memory volume alone does not explain the gap.   We also report what did not work: an unmitigated 4-bit quality cost, roughly half the convolution channels ending up inert and impossible to remove, and a vocabulary larger than this model size warrants.", "5": "2026-08-22T19:13:00.448837"}
{"0": 19, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20204v1", "3": "ContractScrub: A benchmark for final review of legal contracts", "4": "ContractScrub: A benchmark for final review of legal contracts. Authors: Yejin Bang, Kirsty Fielding, Brandan Oliver, Brian Birke, Nabeel Seedat. Categories: cs.AI cs.CL. Legal work, with its heavy reliance on processing large amounts of text, is often considered one of the domains most exposed to the use of LLMs. Contract ``scrubbing,'' the final review of transactional agreements for errors and inconsistencies, is a particularly suitable task for automation, because it is routine, painstaking work requiring detailed attention to long documents. Scrubbing also seems to align naturally with the general capabilities expected of frontier LLMs around long-context reasoning, consistency checking, and named entity recognition (NER). Despite the economic value and potential for automation, no formal evaluations of LLMs performing contract scrubbing have been conducted. We introduce ContractScrub, the first benchmark designed to evaluate contract scrubbing capabilities, comprising contracts hand-crafted by experienced lawyers over diverse error categories such as misuse of defined terms, incorrect references, and inconsistent language. Frontier models perform surprisingly poorly with only one model reaching 0.75 macro average recall despite strong performance on seemingly related general benchmarks, demonstrating the practical limits of current models and the importance of narrowly targeted, domain-specific benchmarks for measuring real-world impact.", "5": "2026-08-22T19:13:00.454600"}
{"0": 20, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20202v1", "3": "MemTrapBench: Benchmarking Cognitive Traps in LLM Memory Use", "4": "MemTrapBench: Benchmarking Cognitive Traps in LLM Memory Use. Authors: Mengru Wang, Haozhe Luo, Zhenqian Xu, Zhixiang Cui, Haoming Xu. Categories: cs.AI cs.CL cs.CY cs.DB cs.LG. Memory has become a key component of large language models, enabling them to retain information and learn from long-term interactions. However, existing memory benchmarks mainly evaluate whether information is correctly extracted, stored, and retrieved, while largely overlooking how retrieved memories reshape model reasoning and affect performance on the current task. We identify memory-induced cognitive traps: even faithfully recorded and semantically relevant memories can distort model reasoning or beliefs and degrade current task performance. To systematically evaluate these failure modes, we introduce MemTrapBench, which covers two forms of cognitive traps: Reasoning Fixation and Belief Distortion. Experiments across two model families and five representative memory frameworks show that MemTrapBench is challenging: all evaluated memory strategies underperform the no-memory setting, with even the strongest methods suffering drops of more than 10%. To mitigate these cognitive traps, we propose AdaptiveMem, a simple yet effective inference-time method that instructs LLMs to avoid memory traps. AdaptiveMem mitigates cognitive traps on MemTrapBench while preserving or improving performance on standard memory benchmarks across diverse memory frameworks.", "5": "2026-08-22T19:13:00.460599"}
{"0": 21, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20201v1", "3": "The Third Restructuring of Software Form: From the Three-Tier Architecture to Storage, Models, and Agents", "4": "The Third Restructuring of Software Form: From the Three-Tier Architecture to Storage, Models, and Agents. Authors: Wei Lin, Tao Zhou, Zhaofei Xie, Changgui Hong. Categories: cs.AI cs.SE. Software form has undergone two paradigm shifts since its inception: Software 1.0, in which instructions determine behavior, and Software 2.0, in which data determines behavior (machine learning). This paper argues that a third shift - Software 3.0, in which context and reasoning determine behavior - is now underway, and contends that its terminal form converges to three elements: a generalized database (the unified abstraction of all persistent state and memory), a large model (the intelligence core that performs reasoning and generation), and an agent (the execution loop connecting the first two). The core argument is as follows: in the traditional three-tier architecture, the user-interface layer will be absorbed by the model's ability to generate interfaces on demand, the business-logic layer will be re-partitioned along \"expressibility x criticality\" into model reasoning and storage constraints (with residual deterministic logic retained as tools), and only the data layer will be elevated into the sole persistent infrastructure. We formalize this convergence thesis, present a minimal reference architecture, report evidence from real prototypes and a live model, and systematically analyze both the conditions under which it holds and the boundaries where it fails - determinism, cost, security, and verifiability delimit the thesis's domain of applicability. We argue that the thesis holds in task domains that are expressible, verifiable, externally stateful, and tool-complete, and that it will reshape the roles of developers, the database industry, and the software-engineering discipline.", "5": "2026-08-22T19:13:00.466731"}
{"0": 22, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20195v1", "3": "From Agent Behaviour to Agent-Friendly Documentation: An Empirical Study of How Coding Agents Discover, Read, and Write Technical Documentation", "4": "From Agent Behaviour to Agent-Friendly Documentation: An Empirical Study of How Coding Agents Discover, Read, and Write Technical Documentation. Authors: Zhijun Gao, Jing Chen. Categories: cs.SE cs.AI cs.HC. Technical documentation is written for human developers, but an increasing share of software changes is now authored by autonomous coding agents. Which documents they consult, when, and what follows remain unknown. We conduct a behaviour-grounded study of agent-documentation interaction across two public datasets: 557 agentic coding sessions from SWE-chat, yielding 94,813 development events including 3,033 documentation interactions; and 33,097 agentic pull requests from AIDev, with 690,260 classified file-level change records. Four findings challenge current documentation practice. First, agents' documentation work is dominated by agent-facing artefacts: instruction files and working notes account for 60.5% of all documentation interactions, versus 10.6% for classical technical documentation and 1.3% for API references. Second, the link between consultation and code editing is unresolved: the adjacent transition probability is 0.002 and the unadjusted three-event lift 1.05, whereas a stage-adjusted model places it above unity (OR 1.33 [1.09, 1.62]); documentation creation is elevated unadjusted (lift 1.67) but its adjusted interval includes unity. Third, no explicit documentation-based validation sequence was observed, and consultation is associated with less immediate testing (lift 0.23, cluster CI 0.08-0.45; adjusted OR 0.39 [0.25, 0.60]). Fourth, consultation is self-initiated (70.2%) far more often than failure-driven (7.5%), and documentation trails code: among multi-commit pull requests changing both, code is touched first 4.7x more often. From these traces we derive a descriptive model of agent-documentation interaction as a two-lobed cycle rather than a linear journey, and show that two widely assumed properties of \"agent-friendly\" documentation - actionability and verifiability - lack consistent behavioural support. We release our pipeline, coding scheme, and event-level data.", "5": "2026-08-22T19:13:00.471255"}
{"0": 23, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20187v1", "3": "Multi-Method Causal Evidence Synthesis: Ranking Candidate Drivers by Convergent Cross-Method Evidence from Observational Data", "4": "Multi-Method Causal Evidence Synthesis: Ranking Candidate Drivers by Convergent Cross-Method Evidence from Observational Data. Authors: Manish Gupta, Dipanjan De. Categories: stat.ME cs.AI. Practitioners inferring causality from observational data usually rely on a single method and treat its output as causal truth. Recent tools select an optimal method for a dataset, and recent ensembles aggregate multiple causal-discovery algorithms into one graph, but little work pools evidence across different mathematical traditions, including non-causal ones. We present Multi-Method Causal Evidence Synthesis (MCES), a framework that ranks which candidate drivers in an observational system are most likely relevant to a set of outcomes, and with what strength of evidence. MCES runs eleven methods across eight mathematical traditions on observational panel data and pools their outputs into a Convergent Evidence Score (CES), a linear opinion pool. CES quantifies convergence of evidence across analytical lenses: the degree to which methods with different assumptions point to the same driver-outcome relationship. It does not claim causal identification in the interventionist sense; it supports hypothesis prioritization, not a transferable probability of causation. MCES first applies Structural-Behavioral Decomposition to remove definitional (algebraic) relationships, then runs all methods, normalizes outputs to [0,1], and pools them. We distinguish MCES from method selection, structural ensembles, prediction ensembles, and literature synthesis. Using synthetic data with embedded ground truth, the Sachs protein-signaling benchmark, six Bayesian-network structure benchmarks, and two further synthetic domains, we show MCES ranks true edges near the top (Precision@5 = 1.0, Precision@10 = 0.96 on the primary scenario), with a low empirical rate of null pairs reaching Moderate-or-higher convergence. Our central point is not that the pool beats every individual method, but that no single method is uniformly best across the evaluated scenarios, so MCES offers a method-agnostic default.", "5": "2026-08-22T19:13:00.479046"}
{"0": 24, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20181v1", "3": "A Standardized Framework for Machine Learning in Power System Protection", "4": "A Standardized Framework for Machine Learning in Power System Protection. Authors: Julian Oelhaf, Georg Kordowich, Paula Andrea P\u00e9rez-Toro, Christian Bergler, Johann J\u00e4ger. Categories: cs.LG cs.AI eess.SP. Studies of machine-learning-based power-system protection increasingly report near-perfect scores, yet the meaning of those scores depends strongly on the evaluation setting. Protection task, physical scope, measurements, timing, targets, preprocessing, and validation often vary jointly and remain incompletely specified. This paper proposes a standardization-oriented framework that treats evaluation design as part of the scientific contribution. It defines seven required study dimensions: protection objective, physical scope, observability, timing and decision windows, targets and sample validity, validation protocol, and evaluation outputs. The framework is instantiated in a bounded case study on the public PROTECT-90 electromagnetic-transient benchmark, comprising 9022 simulated episodes from a 90 kV double-line topology, for onset-conditioned fault classification and localization. Under centralized sensing, simulation-metadata-aligned 20 ms windows, and episode-grouped validation, a multi-layer perceptron (MLP) achieved a five-fold mean macro-averaged F1 score of 0.991 +/- 0.001 for classification and a localization mean absolute error of 10.20 +/- 0.25% of line length (mean +/- std across episode-grouped folds). Extending the decision horizon to 50 ms preserved this task-dependent performance asymmetry, while reduced observability approximately doubled the MLP localization error but had little effect on classification. A synchronized two-ended conventional locator outperformed the learning locators under its richer clean information set, and measurement degradation showed that clean predictive performance did not determine robustness. The framework turns evaluation assumptions into explicit, reproducible evidence and provides a basis for more comparable, auditable evaluation and future certification-oriented assessment of machine-learning protection functions.", "5": "2026-08-22T19:13:00.484704"}
{"0": 25, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20169v1", "3": "Task-CoEvolve: Efficient Harness Optimization via Adaptive Validation Task Selection", "4": "Task-CoEvolve: Efficient Harness Optimization via Adaptive Validation Task Selection. Authors: Atsuyuki Miyai, Kiyoharu Aizawa, Toshihiko Yamasaki. Categories: cs.CL cs.AI cs.LG. We present a novel approach to efficient LLM agent harness optimization through adaptive validation task selection. Harness optimization iteratively rewrites the harness code based on validation performance, enabling substantial performance gains without updating the underlying model weights. Existing approaches, however, evaluate a fixed validation set in full at every iteration, incurring substantial evaluation costs even on tasks that become less discriminative as the harness evolves. We propose $\\textbf{Task-CoEvolve}$, which co-evolves the validation tasks with the harness by addressing two challenges: selecting informative tasks and estimating full-set performance from partial evaluations. Task-CoEvolve builds on the observation that tasks on which candidate harnesses disagree are more informative for distinguishing among them than tasks that are consistently solved or failed. It uses variance-weighted sampling based on past outcomes to focus evaluation on tasks near the agent's capability frontier, with the sampling distribution adapting as the harness evolves. It then estimates full-set scores from the sampled tasks by accounting for their sampling probabilities, enabling consistent comparisons across iterations despite evaluating different subsets. Experiments on online text classification and Terminal-Bench 2.1 show that Task-CoEvolve consistently outperforms fixed-subset baselines and matches the final performance of full-set search while reducing the number of evaluations during optimization by 80%. Code will be released at https://github.com/Agent4Science-UTokyo/Task-CoEvolve.", "5": "2026-08-22T19:13:00.493276"}
{"0": 26, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20161v1", "3": "DARS: Dual-Level Credit Assignment RL with Structured Reasoning for Instruction-Based Image Editing", "4": "DARS: Dual-Level Credit Assignment RL with Structured Reasoning for Instruction-Based Image Editing. Authors: Haoxiang Cao, Jiajiong Cao, Xuanpu Zhang, Changqian Yu, Chaoqun Wang. Categories: cs.AI. Instruction-based image editing uses a planner-renderer pipeline: a vision-language model (VLM) first converts the instruction into an edit plan, and a diffusion model then executes that plan. Training such systems with only final-image rewards is inefficient because a poor edit does not reveal whether additional optimization should place more emphasis on the planner or the renderer, and even planner-dominant cases remain difficult to localize within a free-form reasoning trace. We present DARS, a reinforcement learning framework for dual-level credit assignment in this two-stage setting. Across modules, multi-plan multi-render rollouts estimate between-plan and within-plan reward variability for soft module routing, while rollout mean rewards provide hardness estimates for an adaptive curriculum. Within the planner, a four-field structured reasoning output enables a prefix-gated reward and token-level advantage reweighting, turning outcome-level feedback into localized supervision. Experiments on five benchmarks show that DARS outperforms a Joint~RL baseline with the same backbone, data, reward model, and rollout budget, with the largest gains on reasoning-intensive edits.", "5": "2026-08-22T19:13:00.499239"}
{"0": 27, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20114v1", "3": "DECOWAM: Decoupled Whole-Body World-Action Model for Legged Mobile Manipulation", "4": "DECOWAM: Decoupled Whole-Body World-Action Model for Legged Mobile Manipulation. Authors: Siyuan Ma, Boshi Zhang, Yutian Zhang, Qinglian Wu, Jiaqi Zhai. Categories: cs.AI cs.RO. Mobile manipulation requires a robot to predict how locomotion and arm motion jointly alter future observations and control. Existing world-action models, developed largely for fixed-base platforms, do not explicitly distinguish camera ego-motion from base and arm actions. Here we introduce DECOWAM, a whole-body world-action model that separates these factors through dedicated conditional interfaces. DECOWAM freezes an adapted FastWAM backbone and trains residual adapters, an action-equivalent future bottleneck distilled from privileged observations, adversarially separated base and arm latents, and base-velocity conditioning for video prediction. We further introduce ARMDOG, a real-robot dataset that synchronizes video, whole-body state and action, and language. On a fixed replay protocol, DECOWAM improved both future-video and action prediction over FastWAM, reducing action MSE by 21.7% with 25.95M trainable adaptation parameters. Across 79 closed-loop trials per method, it achieved the highest observed whole-body coordination and base-displacement robustness among the compared systems, while task completion remained comparable to the strongest baseline. These results show that embodiment-aware factorization can support parameter-efficient joint visual prediction and whole-body control under moving viewpoints.", "5": "2026-08-22T19:13:00.506239"}
{"0": 28, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20104v1", "3": "Structured Affinity for Unsupervised Visual Class-Incremental Memory in Deep Artificial Immune Networks", "4": "Structured Affinity for Unsupervised Visual Class-Incremental Memory in Deep Artificial Immune Networks. Authors: Siphesihle Sithungu. Categories: cs.CV cs.AI cs.LG. Artificial immune networks (AINs) are naturally memory-forming systems, but conventional visual AINs often rely on flattened vector affinity that ignores spatial structure. This paper studies whether structured, gradient-free immune affinity can make Deep AINs viable as replay-free visual class-incremental representation-memory learners. Visual B-cells are formalized as structured templates, including shifted-template affinity, zero-normalized cross-correlation (ZNCC) filters, and feature-map binding profiles. A repertoire is treated both as memory and as a representation-inducing basis, while depth is obtained by passing binding-profile response maps to subsequent immune layers. The resulting Deep AIN exhibits adaptive latent coordinate reorganization: as new classes arrive, the binding-profile space evolves while retaining recoverable structure for earlier classes. Experiments on sklearn digits, MNIST, Fashion-MNIST, and KMNIST show that preserving response maps is critical. Scalar binding-profile variants underperform, whereas feature-map Deep AINs learn class-discriminative visual memory without replay, label-driven immune updates, or backpropagation through the immune layers. On sklearn digits, downstream probes fitted on the learned binding profiles reach 0.939 final balanced accuracy with logistic regression and 0.902 with 1-nearest-neighbour after all ten classes are encountered, with initial-class retention of 0.978. Adaptive layer-wise scale calibration further improves the two-layer feature-map Deep AIN to 0.978 balanced accuracy. With the same calibration rule, Fashion-MNIST reaches 0.814 and KMNIST reaches 0.853. These probes are external validation tools, not components of the AIN. The results identify structured affinity, response-map preservation, adaptive latent reorganization, and layer-wise scale calibration as key mechanisms for replay-free visual immune memory.", "5": "2026-08-22T19:13:00.512517"}
{"0": 29, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20087v1", "3": "Towards Professional Tennis Styles for Humanoid Robots with Adaptive Motion Planning and Tracking", "4": "Towards Professional Tennis Styles for Humanoid Robots with Adaptive Motion Planning and Tracking. Authors: Tao Huang, Ruofei Liu, Xuchen Tang, Xinyin Zhang, Junli Ren. Categories: cs.RO cs.AI. Humanoid robots have recently demonstrated promising capabilities in real-world ball sports. However, achieving professional motion styles while maintaining strong task performance remains challenging. In this work, we propose AdaPT, an Adaptive Motion Planning and Tracking framework that learns professional tennis serving and rally styles directly from broadcast videos. This hierarchical design is motivated by the key insight that the planner generates stylistic kinematic motions, while the tracker executes them with minimal interference with planning. Despite its effectiveness in simulation, a substantial sim-to-real gap emerges: tracking performance inevitably degrades on real robots, and this degradation is partially overlooked by autoregressive planning and further compounded by noisy perception. To address these issues, our adaptation mechanism improves tracking robustness by learning to track randomized execution speeds, while conditioning the planner on a learned motion-speed adapter to mitigate compounding errors. Real-world experiments on the Unitree G1 demonstrate the effectiveness of our adaptation mechanism in bridging the sim-to-real gap. We further deploy AdaPT policies on the full-size Dobot Atom humanoid robot (1.7m) and demonstrate in-the-wild serving without motion capture. Beyond these results, our real-world experiments reveal both algorithmic and engineering insights for future humanoid ball-sports systems. Videos and code are available on our \\href{https://humanoidtennis.github.io/AdaPT/}{project website}.", "5": "2026-08-22T19:13:00.517030"}
{"0": 30, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20084v1", "3": "Evidence-Gated Task and Motion Planning with Vision-Language Models", "4": "Evidence-Gated Task and Motion Planning with Vision-Language Models. Authors: Tsunehiko Tanaka, Matthew Stephenson, Alistair Macvicar, Edgar Simo-Serra. Categories: cs.RO cs.AI. Robots executing long-horizon manipulation tasks from natural-language instructions must reason about both semantic task structure and geometric feasibility. However, under partial observability, the availability of goal-relevant objects may be uncertain. In such cases, approaches that combine Vision-Language Models (VLMs) with Task and Motion Planning (TAMP) may generate subgoals that rely on the VLM's prior knowledge without observational support, leading to execution failures or unintended outcomes. We propose Evidence Acquisition and Feasibility Gating (EAFG), a framework that acquires visual evidence through VLM-generated exploratory subgoals and TAMP-based execution. EAFG then applies a feasibility gate to decide whether to proceed with task planning, acquire further evidence, or halt. Our experiments show that, in cooking tasks with ambiguous object use, EAFG improves recipe completion by discovering task-relevant objects before planning. For instructions requiring an absent object, EAFG promotes appropriate halt decisions and reduces repeated attempts to manipulate that object.", "5": "2026-08-22T19:13:00.524847"}
{"0": 31, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20338v1", "3": "ConceptGuard: Benchmarking Context-Sensitive Unlearning in Large Language Models", "4": "ConceptGuard: Benchmarking Context-Sensitive Unlearning in Large Language Models. Authors: Sahil Kale, Ian Harris. Categories: cs.CL. Large Language Models (LLMs) increasingly require selective removal of harmful or sensitive knowledge, called unlearning, yet existing methods and benchmarks fail to evaluate this capability completely. Current approaches rely on disjoint forget and retain sets composed of independent facts, and measure success using simple and direct factual recall. This framing fails to capture a key requirement of unlearning, namely the ability to eliminate harmful behaviors while preserving benign and beneficial knowledge. We argue that effective unlearning must operate at the level of concepts, ensuring complete removal of unsafe applications while maintaining their correct and useful usage, thereby achieving conceptually meaningful and complete unlearning. To better evaluate unlearning techniques from such a practical viewpoint, we introduce the notion of dual-use concepts: concepts that can be used in both harmful and benign contexts. Building on these concepts, we construct a benchmark called ConceptGuard where forget and retain sets are explicitly complementary in concept usage. Our benchmark uniquely enables unlearning to be explored and gauged at the level of concepts, instead of sparse facts, and evaluation is intent-sensitive with the goal of maximizing contextual separation to promote safer behavior. We demonstrate that current unlearning techniques perform poorly under this setting, showing weak contextual separation alongside poor performance in ROUGE and concept-level metrics. Our results reveal strong forgetting-utility trade-offs, limited gains in contextual sensitivity, and poor consistency in concept-level control across methods, and provide ideas for unlearning approaches that better align with real-world safety requirements. Our dataset is publicly available.", "5": "2026-08-22T19:13:04.201833"}
{"0": 43, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20153v1", "3": "FormalTCS: Benchmarking End-to-End Frontier Formal Theoretical Computer Science Research of Large Language Models", "4": "FormalTCS: Benchmarking End-to-End Frontier Formal Theoretical Computer Science Research of Large Language Models. Authors: Dingzirui Wang, Xuanliang Zhang, Keyan Xu, Qingfu Zhu, Wanxiang Che. Categories: cs.CL. Large language models (LLMs) have shown growing potential for automated theoretical computer science (TCS) research, yet existing benchmarks remain far from realistic research settings. We introduce \\ourbenchmark, an expert-validated benchmark for evaluating LLMs on frontier, end-to-end TCS research. \\ourbenchmark contains $175$ instances drawn from papers accepted to STOC, FOCS, SODA, and COLT in 2025-2026, preserving paper-specific definitions, assumptions, and proof dependencies, with expert-verified Lean formalizations and proofs. Evaluations of leading LLMs reveal that current models remain far from reliably completing the full research pipeline. In particular, autoformalization is the sharpest bottleneck: the best model achieves only $11.5$ on translating natural-language claims into formal theorem statements, compared with $28.6$ Pass@8 when proving human-provided formal statements. Building on \\ourbenchmark, we further develop an automated TCS research framework that generates, formalizes, filters, and proves new claims. Of $64$ generated claims, only $6$ ultimately pass expert evaluation and proof verification, indicating that beyond formalization, limited research taste remains another major barrier to autonomous TCS research.", "5": "2026-08-22T19:13:04.274489"}
{"0": 44, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20129v1", "3": "Multi-Agent Orchestration with the Common-Sense Reasoning Capabilities of LLMs for Autonomous Driving", "4": "Multi-Agent Orchestration with the Common-Sense Reasoning Capabilities of LLMs for Autonomous Driving. Authors: Mehdi Azarafza, Faezeh Pasandideh, Ali Ehteshami Bejnordi, Stefan Henkler, Achim Rettberg. Categories: cs.MA cs.CL cs.CV. Autonomous vehicles require robust perception and decision-making capabilities to operate in diverse and unseen scenarios. While reinforcement learning and rule-based methods can provide effective control and safety mechanisms, their performance may degrade in situations requiring contextual reasoning. Large Language Models (LLMs) have demonstrated strong capabilities in understanding multimodal information and generating contextual reasoning, however, their use for direct vehicle control can introduce latency and hallucination risks. To address these limitations, a hybrid framework is proposed. This system uses an orchestrator to coordinate PPO-trained reinforcement learning and PID control, with LLM common-sense reasoning applied throughout the framework. LLM reasoning is further employed iteratively to refine the RL reward function for dynamic driving environments. The proposed framework is evaluated in highly randomized CARLA scenarios under diverse environmental and traffic conditions. The results demonstrate the potential of integrating LLM-based reasoning with conventional autonomous driving methods while retaining structured control and safety mechanism.", "5": "2026-08-22T19:13:04.282357"}
{"0": 45, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20116v1", "3": "When Text and Numbers Disagree: Evidence Arbitration in Large Language Models", "4": "When Text and Numbers Disagree: Evidence Arbitration in Large Language Models. Authors: Mattia Carletti, Edward Phillips, Fredrik K. Gustafsson, Patitapaban Palo, Lei Clifton. Categories: cs.CL. Large language models (LLMs) are increasingly used in settings where textual summaries, numerical observations, and external tool outputs may provide conflicting evidence. We study how LLMs arbitrate between such sources when they support opposing decisions. To do so, we introduce a controlled synthetic benchmark in which latent risk trajectories generate both numerical time series and natural language summaries, allowing us to construct conflicts where exactly one evidence source is aligned with the ground-truth label. This design lets us independently manipulate modality, temporal recency, source reliability, and evidence provenance. Across open-weight instruction-tuned models, we find that arbitration behaviour is systematic rather than random: models exhibit distinct text-versus-number preferences, follow temporal recency more consistently than explicit reliability cues, and can over-rely on external forecasts even when they conflict with direct contextual evidence. These results suggest that current LLMs often rely on heuristic arbitration strategies when integrating heterogeneous evidence, highlighting a failure mode for tool-augmented decision systems.", "5": "2026-08-22T19:13:04.288058"}
{"0": 46, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20106v1", "3": "OenoBench: A Wine-Domain Benchmark for Knowledge-Grounded Evaluation of Large Language Models", "4": "OenoBench: A Wine-Domain Benchmark for Knowledge-Grounded Evaluation of Large Language Models. Authors: Nikita Khudov. Categories: cs.CL. We introduce OenoBench, a wine-domain knowledge benchmark of 3,266 multiple-choice questions across six pillars (regions, grape varieties, viticulture, winemaking, producers, business) and four difficulty tiers. The corpus is built from 38,104 atomic, source-anchored facts extracted by 35 provenance-verified scrapers from government registries (INAO, TTB, OIV), peer-reviewed journals, and Wikipedia/Wikidata. Our methodological contribution is an LLM-driven pipeline in which language models reformat verified facts and audit the result, but never serve as the source of truth: every claim traces to a URL, every question is generated by one of five strategies across five generator families, and every question is scored by a nine-agent audit calibrated against a human gold sheet via Cohen's $\u03ba$. Evaluating sixteen frontier configurations, we find: (i) overall accuracy spans 53%-84%, led by o3 at 83.6%; (ii) reasoning-mode lift concentrates in DeepSeek R1 (+6.8pp) and is absent in Claude Opus and Gemini Pro; (iii) Anthropic shows +9pp self preference on its own questions while Google shows -8pp inverse preference; (iv) frontier open-weight models share the cost-vs-accuracy Pareto frontier with proprietary reasoning models; and (v) every config gains around 33pp on closed-book solvable items, revealing a parametric-recall ceiling that only the contextual slice avoids. We release corpus, audit findings, human-review app, and construction code under CC-BY-SA-4.0.", "5": "2026-08-22T19:13:04.294628"}
{"0": 47, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20099v1", "3": "Reward-Guided Autoregressive Graph Generation for Efficient Multi-Agent Communication Topology Design", "4": "Reward-Guided Autoregressive Graph Generation for Efficient Multi-Agent Communication Topology Design. Authors: Poomphob Suwannapichat, Boonyarit Changaival, Caesar Wu, Pascal Bouvry. Categories: cs.MA cs.CL cs.LG. LLM-based Multi-Agent Systems (MAS) achieve strong performance on complex reasoning tasks by coordinating multiple agents, but at the cost of substantial token consumption. Recent work on automatic topology design, ARG-Designer, has reframed this problem as autoregressive graph generation. However, its training objective provides no explicit incentive for the model to generate sparse and efficient topologies. We address this limitation by introducing a Reward-Guided Autoregressive Graph Generation (RGA-Designer) inspired by Reinforcement Learning from Human Feedback (RLHF). We train a reward model that jointly captures task correctness and structural compactness, and then fine-tune the pretrained graph generator using the reward model as feedback. Our method preserves task accuracy at the level of ARG-Designer while reducing token consumption by an average of 20.5%.", "5": "2026-08-22T19:13:04.301004"}
{"0": 48, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20083v1", "3": "SABET-QA: Temporal Knowledge Graph Question Answering", "4": "SABET-QA: Temporal Knowledge Graph Question Answering. Authors: Brahim Touayouch, Mirette Moawad, Dmitry Akulov. Categories: cs.CL cs.AI. Question Answering over Temporal Knowledge Graphs (TKGQA) requires reasoning over time-sensitive facts, yet existing embedding-based methods struggle with multi-step queries due to single-pass reasoning pipelines. We propose SABET-QA, a framework that iteratively refines reasoning states across multiple hops via a bidirectional entity-temporal scoring mechanism and a slot-aware contextualization module that aligns question semantics with temporal KG embeddings. A differentiable working memory enables progressive hypothesis refinement, while auxiliary temporal boundaries serve as coarse supervision when available. Experiments on CronQuestions, Complex-CronQuestions, MultiTQ, and TimeQuestions demonstrate consistent improvements over strong baselines, particularly on complex multi-step temporal queries.", "5": "2026-08-22T19:13:04.306459"}
{"0": 49, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20061v1", "3": "Let's Scale Step by Step: Compute-Efficient Hyperparameter Transfer for Large-Scale Mixture-of-Experts", "4": "Let's Scale Step by Step: Compute-Efficient Hyperparameter Transfer for Large-Scale Mixture-of-Experts. Authors: Nayeon Kim, Hojin Lee, Yunju Bak, Jaesun Park, Boseop Kim. Categories: cs.LG cs.AI cs.CL. Mixture-of-Experts (MoE) architectures significantly expand model capacity without a proportional increase in computational cost. However, optimizing their hyperparameters---particularly the learning rate---at extreme scales of both model size and token budget via sweeping remains computationally prohibitive. In this paper, we propose a compute-efficient, two-step hyperparameter transfer framework that estimates optimal learning rates for training large MoE models by transferring them across scaling model widths, and subsequently extrapolating to trillion-token horizons. First, we formulate a Maximal Update Parameterization ($\u03bc$P) adaptation for MoE architectures utilizing Multi-head Latent Attention (MLA) and the Muon optimizer, demonstrating that optimal learning rates transfer consistently across width-scaled models. Second, we extend this transferability along the token dimension by establishing a predictive scaling law. By applying linear regression to the optimal values derived from small proxy models on limited budgets, we successfully extrapolate the ideal learning rate to massive training horizons (e.g., 10 trillion tokens) with high fidelity ($R^2=0.95$). Consequently, this indicates that proxy training on small models is sufficient to determine the optimal learning rate for the extensive training of large-scale MoEs. We apply the proposed methodology to pretrain our foundation model (155B total, 17B active parameters) from scratch, and the stable training and evaluation results validate that optimal configurations for full-scale target models can be accurately predicted with minimal ablation costs.", "5": "2026-08-22T19:13:04.314859"}
{"0": 50, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20047v1", "3": "Auditing Cross-Lingual Fairness in Language Model Watermarking", "4": "Auditing Cross-Lingual Fairness in Language Model Watermarking. Authors: Alexander Nemecek, Osama Zafar, Debargha Ganguly, Vikash Singh, Vipin Chaudhary. Categories: cs.CL cs.CR cs.LG. Watermarking schemes for large language model output are evaluated almost exclusively on English text using each scheme's detection threshold and a narrow set of quality measurements. Multilingual deployment exposes evaluation-design choices that are inconsequential on English but determine conclusions cross-lingually. We propose an evaluation framework with four components: detection thresholds calibrated empirically per deployment context, a threshold-independent companion measurement that distinguishes calibration failures from detection failures, three disjoint quality measurement paradigms (distributional, paired-semantic, and reference-perplexity), and a generalized-entropy decomposition of cross-language disparity over a typological family partition. Applied to six watermarking schemes, three open-weight generators, eleven languages spanning four scripts and eight typological families, and both base and instruction-tuned regimes, the framework reveals failure modes that single-language single-paradigm evaluation cannot surface. Across detection and quality, observed disparity is predominantly between-family on the typological partition, indicating that cross-lingual fairness gaps in watermarking are structural to language properties rather than idiosyncratic to particular languages.", "5": "2026-08-22T19:13:04.319698"}
{"0": 51, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19981v1", "3": "HealMed: Multilingual Evaluation of Large Language Models in Medicine", "4": "HealMed: Multilingual Evaluation of Large Language Models in Medicine. Authors: Yingjian Chen, Fan Gao, Sherry T. Tong, Haoyu Zhang, Aosong Feng. Categories: cs.CL. We present HealMed, an expert-reviewed benchmark for multilingual evaluation of large language models in medicine. HealMed contains 1,000 examples in each of nine languages, drawn from nine datasets and covering three task formats: MCQA, NLI and open-ended QA. The benchmark was developed over two years by 23 physicians and medical experts based across nine countries and regions. Each translation was evaluated and revised by two experts fluent in English and the corresponding target language. On HealMed, performance declined most in low-resource languages, although the size of the gap varied markedly across languages and models. The strongest proprietary models were the most stable across languages, whereas many open-source and medically specialized models showed larger and less consistent gaps. Medical specialization alone did not ensure multilingual robustness. Furthermore, expert revision could either raise or lower measured performance, indicating that translation quality materially affects cross-language evaluation results.", "5": "2026-08-22T19:13:04.328837"}
{"0": 52, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19971v1", "3": "Robust Incomplete Multimodal Sentiment Analysis via Iterative Proxy Correction", "4": "Robust Incomplete Multimodal Sentiment Analysis via Iterative Proxy Correction. Authors: Zhifa Geng, Subin Huang, Hao Guo, Junjie Chen, Sanmin Liu. Categories: cs.CL. Multimodal sentiment analysis aims to infer affective states by integrating language, visual, and acoustic cues. However, real-world multimodal inputs are often incomplete or corrupted, which can weaken cross-modal complementarity and introduce misleading information into downstream fusion. Existing proxy-based methods for incomplete MSA commonly rely on one-shot proxy construction to compensate for degraded language information, but the generated proxy may be coarse or unreliable at initialization. Prematurely injecting such a proxy into multimodal reasoning can propagate initial errors and compromise sentiment prediction. To address this limitation, we propose an iterative proxy correction framework for robust incomplete MSA. Our method constructs a language-oriented proxy from non-language modalities and progressively refines it under multimodal context through gated residual correction. The corrected proxy is then adaptively fused with the observed language representation according to an estimated language reliability score, allowing the model to balance proxy-based compensation and trustworthy linguistic evidence. In addition, we introduce a stage-wise latent correction objective that uses the complete language representation as a training-time semantic anchor to stabilize the proxy refinement trajectory. Extensive experiments on MOSI, MOSEI, and SIMS under diverse missing-modality settings demonstrate that the proposed framework consistently outperforms competitive baselines and achieves robust sentiment prediction under incomplete inputs.", "5": "2026-08-22T19:13:04.335264"}
{"0": 53, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19957v1", "3": "Natural Language Code Retrieval for 1C:Enterprise: An Open Benchmark and Efficient Bi-Encoder", "4": "Natural Language Code Retrieval for 1C:Enterprise: An Open Benchmark and Efficient Bi-Encoder. Authors: Konstantin Chesnokov, Chingiz Mingazov. Categories: cs.CL. Natural language code retrieval is a rapidly evolving task in computer science. However, the 1C:Enterprise ecosystem combines Russian syntax with highly domain-specific terminology, for which open datasets and specialized models have been virtually non-existent. We present a comprehensive pipeline for 1C code retrieval: an open benchmark of 3,413 real-world, PII-scrubbed query-code pairs, a reproducible evaluation harness, and a specialized bi-encoder. To overcome scarce labeled data, we fine-tune on 784,057 synthetic triplets generated by google/gemma-4-26B-A4B-it from public code repositories, using Matryoshka Representation Learning (MRL) and a privacy-aware tokenizer. Because the benchmark subsets differ in size, we report balanced-subset macro, query-weighted micro, and forum-only results. Our model reaches 0.5992 balanced macro nDCG@10, 0.5044 micro, and 0.4617 on forum, versus 0.4932 macro for the baseline architecture and 0.5404 for google/embeddinggemma-300m. Removing every benchmark example flagged by the conservative exact/13-gram overlap audit leaves 0.6011 balanced macro (0.5010 micro), indicating that detected train-benchmark overlap does not explain the headline result. MRL truncation to 256 dimensions preserves 99.9% of retrieval quality while reducing dense-index storage and exact similarity arithmetic by a factor of three.", "5": "2026-08-22T19:13:04.345013"}
{"0": 54, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19942v1", "3": "Dynamic Gated Cross-Modal Fusion with Sarcastic-aware Contrastive Regularization for Multimodal Sarcasm Detection", "4": "Dynamic Gated Cross-Modal Fusion with Sarcastic-aware Contrastive Regularization for Multimodal Sarcasm Detection. Authors: Hao Guo, Subin Huang, Junjie Chen, Zhifa Geng, Sanmin Liu. Categories: cs.CL. Multimodal sarcasm detection aims to identify sarcastic intent from multimodal content, where inconsistencies between literal meaning and contextual cues often signal irony. This task has attracted increasing research attention. However, accurate detection remains challenging due to instance-dependent modality contributions and misleading semantic consistency, where surface-level alignment masks underlying contradictory intent. Existing methods often rely on fixed fusion strategies and treat sarcasm as generic cross-modal mismatch, limiting their ability to capture subtle sarcasm cues and instance-specific modality interactions. To address these challenges, we propose a novel MSD framework that integrates Dynamic Gated Cross-Modal Fusion with Sarcastic-aware Contrastive Regularization (SaCR). Specifically, a bidirectional gated interaction module performs cross-modal feature filtering and adaptively calibrates textual and visual contributions at the instance level. A dynamic fusion gate further balances modality importance to generate more robust multimodal representations. Furthermore, SaCR is introduced as a label-aware contrastive regularization objective that encourages semantic consistency for non-sarcastic samples while suppressing misleading consistency in sarcastic cases. The proposed framework is trained end-to-end with a multi-objective learning strategy that jointly optimizes multimodal classification and auxiliary unimodal supervision. Extensive experiments on MMSD and MMSD2.0 demonstrate that the proposed method consistently outperforms strong baselines.", "5": "2026-08-22T19:13:04.353241"}
{"0": 55, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19920v1", "3": "Learning how to Forget: Fine-tuning for Long-Context Sparse Attention", "4": "Learning how to Forget: Fine-tuning for Long-Context Sparse Attention. Authors: Matthias Seeger, Zeyu Zhang, Vihang Patil, Konstantinos Benidis, Sebastian Schelter. Categories: cs.CL. A lot of prior work addressed key-value (KV) cache selection and compression by sparse attention to enable long-context inference for transformer language models without excessive hardware budgets. We provide a new method for fine-tuning models with sparse attention. It works for any KV cache policy, runs on a moderate hardware budget (e.g., a single Nvidia A100 GPU with 40 GB RAM), and allows the model to co-adapt with the policy, often outperforming models trained with exact attention (sequence parallelism). We also provide an efficient implementation of H2O sparse attention (the leading policy in our experiments) with dedicated scaled dot product attention kernel support. KeysAndValues (https://github.com/awslabs/keys_values), a new open source library for long-context inference and fine-tuning, provides easy-to-use and performant code for all methods discussed here.", "5": "2026-08-22T19:13:04.361220"}
{"0": 56, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19893v1", "3": "Interrupting the Loop: Periodic Subject Changes Raise Judged Surprise and Connection in Base Language Models", "4": "Interrupting the Loop: Periodic Subject Changes Raise Judged Surprise and Connection in Base Language Models. Authors: Roberto I. Ono Filho. Categories: cs.CL cs.AI. Where does the novelty a base language model produces with no task come from, and what can an LLM judge of a long stream actually see? We dismantle a cognitively inspired generation loop over 24 conditions on three base models. Most of its effect lives in one operation: a new subject injected every few hundred tokens (an interruption) into a stream whose literal repetition is damped (habituation). We judge windows of generated text only, with the premise as the unit (n=10) and a judge measured for repeatability, against a second judge family and against human readers. Under that protocol the interruption raises judged surprise by 1.2 to 1.4 points and connection by 0.8 over habituation alone. A connective that asks for continuity hurts; a bare paragraph break adds nothing detectable on fresh text; a reset context does at least as well as a kept one; and a pre-registered replication on new premises confirms the primary contrast. Three things the window judge could not see changed the first version of this study, and we think they are of general use. The judge scores the experimenter's injected sentence as the model's own. A fixed rotation of injected sentences makes the model replay its earlier segments from beyond the judge's horizon, and the judge scores the replay as surprise and connection (65-80% of post-interruption windows at periods 150-300). And the local gains do not compose: no arm produces an integrated document. The salience monitor, the in-loop judge, memory across interruptions and a judge-gated Review run with a gate that opens add nothing. On a problem with a verifier (online bin packing), the interruption multiplies valid, distinct candidate heuristics three- to fourfold without raising the quality of the best. We report an evaluation protocol for long generation and a controlled characterization of a simple intervention, not a mechanism of creativity.", "5": "2026-08-22T19:13:04.368848"}
{"0": 57, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19880v1", "3": "EnvHarness: Awakening Static Worlds for Agent Learning", "4": "EnvHarness: Awakening Static Worlds for Agent Learning. Authors: Chengsong Huang, Zifeng Wang, Rujun Han, Jun Yan, Yanfei Chen. Categories: cs.AI cs.CL cs.LG. LLM agents learn by interacting with environments, yet these environments are hand-built and static: blind to an agent's weaknesses, and quickly left behind as it improves. While recent environment generation methods attempt to address this, they require domain-specific pipelines, rely on expensive or unreliable verifiers, and still produce static environments. To alleviate the engineering burden of rebuilding environments from scratch, we propose Environment Harness (EnvHarness), a programmable layer of plug-in components that wraps a static environment to reshape its behavior without modifying the underlying logic. Operating through standard interfaces, EnvHarness applies across diverse domains while ensuring every reshaped environment retains its original verifier. To automate this process, we introduce EnvRigger, which treats the target policy as a black box, observing its execution trajectories to synthesize EnvHarness components targeting diagnosed flaws, and validating them via fresh rollouts. Across five benchmarks in four domains, EnvHarness outperforms both original environments and domain-specific environment generation pipelines, achieving up to a 9.0-point improvement on held-out instances with 9.8% fewer execution steps. Furthermore, EnvHarness provides a superior optimization signal for reinforcement learning, enabling continuous, targeted co-evolution of the policy and its environment.", "5": "2026-08-22T19:13:04.379182"}
{"0": 58, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19875v1", "3": "A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queries", "4": "A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queries. Authors: Mahyar Abbasian, Saba A. Farahani, Arshia Ilaty, Hung Cao, Ramesh Jain. Categories: cs.CL cs.AI. Patients often submit short, underspecified queries to healthcare chatbots that lack the patient-specific information needed to determine an appropriate response. Although these queries may be linguistically clear, they can support multiple plausible answers depending on undisclosed factors such as symptoms, diagnoses, medications, allergies, or dietary restrictions. A language model answering such a query directly may therefore rely on unsupported assumptions about the patient. We introduce a knowledge-guided agentic framework for mitigating patient-context ambiguity before final response generation. The framework operates between the patient and an otherwise unchanged downstream language model. It interprets the initial query, uses a task-specific knowledge graph to construct a set of plausible hypotheses, identifies the missing patient-context variables needed to distinguish among them, and asks targeted follow-up questions. The original query and the acquired context are then combined into a clarified prompt for the downstream model. We evaluated the framework across five language models using two controlled ambiguity-mitigation benchmarks: diagnosis retrieval from 1,034 symptom queries with clinically relevant evidence systematically masked, and dietary-safety classification from 487 queries with decisive health context omitted. The framework was compared with direct answering of the underspecified query and with rephrasing the same query without acquiring new patient information. In diagnosis retrieval, it increased overall exact Top-1 accuracy by at least 57.1 percentage points and selective exact Recall@5 by at least 77.7 percentage points across the five evaluated models compared with direct prompting. In dietary-safety classification, it improved accuracy across all five models and achieved the highest Matthews correlation coefficient for four...", "5": "2026-08-22T19:13:04.385199"}
{"0": 59, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19861v1", "3": "PolicyGuide: From Guarding One Action to Guiding the Whole Workflow for Policy-Compliant LLM Agents", "4": "PolicyGuide: From Guarding One Action to Guiding the Whole Workflow for Policy-Compliant LLM Agents. Authors: Seongjae Kang, Taehyung Yu, Sung Ju Hwang. Categories: cs.AI cs.CL cs.LG. Customer-service LLM agents must follow organizational policy when acting on a user's behalf. Compliance failures arise from either forbidden actions, such as granting an ineligible change, or omitted procedural requirements, such as identification or confirmation. Runtime safeguards can intervene on risky actions, but action-local checks do not guide an agent through a multi-step procedure. Workflow-following systems support prescribed process execution, but primarily target workflow completion rather than safeguarding agent behavior. PolicyGuide instead compiles each domain policy into a workflow graph and invokes a proactive verifier at user-turn boundaries. From persisted graph state, the verifier reconciles open requests and returns step-specific remediation along a policy-compliant path. Across the $\u03c4^2$-bench airline, retail, and telecom domains with a GPT-5.4 agent and verifier, PolicyGuide raises mean $\\mathrm{Pass}^4$ from $0.42$ to $0.62$, with the largest gain on telecom ($0.19$ to $0.61$), the most workflow-structured domain. The same workflows transfer to Claude Sonnet 4.6 and Gemini 2.5 Pro agents. Complementary evaluations find the lowest observed attack-success rate under adversarial users and the strongest procedural compliance in an author-designed workflow-level validation.", "5": "2026-08-22T19:13:04.393030"}
{"0": 60, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19825v1", "3": "Towards Clinically Faithful Medical Image Captioning via Enhanced Vision-Language Alignment", "4": "Towards Clinically Faithful Medical Image Captioning via Enhanced Vision-Language Alignment. Authors: Yunseo Lee, Hyun Jun Kim, Heeseung Shin, Changwon Lim. Categories: cs.CV cs.CL. Medical image captioning is a technique that accelerates early-stage diagnostic workflows and enhances the interpretability of medical diagnostic AI systems. However, unlike general image captioning, clinically reliable captioning remains challenging due to grayscale-based modalities, subtle anatomical cues, specialized medical phrasing, and variations in data quality. Despite recent advances in large vision-language models, fluent outputs do not necessarily guarantee sufficient alignment with clinical concept spaces or evaluation criteria. To address this issue, we propose a framework that strengthens clinical alignment by separating and enhancing training-time alignment and inference-time alignment. We build a medical image captioning pipeline that integrates single/dual vision encoders based on BioMedCLIP and SigLIP2, a Q-Former, and a LLaMA-based decoder, and examine the contribution of auxiliary learning for UMLS concept/type prediction. At inference, we apply single-embedding-based reranking to select the best caption among candidates, while at training we introduce MedPAIR-SCST, which combines clinically relevant rewards to shift the generative distribution toward improved clinical alignment. Our experiments show that complementary visual representations with a multi-encoder design and concept-level auxiliary learning help preserve clinically meaningful information. Furthermore, inference-time reranking provides a practical way to improve semantic and clinical alignment without additional training, whereas MedPAIR-SCST goes beyond selection by directly improving the model's distribution to generate more consistent and clinically grounded captions. These findings suggest that jointly leveraging selection-based alignment and reinforcement-learning-based alignment can promote more trustworthy medical image captioning even in data-constrained settings.", "5": "2026-08-22T19:13:04.398700"}
{"0": 61, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20337v1", "3": "Information on trajectories: martingales and random times", "4": "Information on trajectories: martingales and random times. Authors: Akshay Balsubramani. Categories: math.PR cs.IT cs.LG math.ST. Accounting for information flow on the path space of trajectories of a nonnegative martingale yields exact variational identities for it, even at arbitrary random times. This recovers the widely used classical concentration inequalities, from Ville to PAC-Bayes, and measures what each one discards. The tail a bound controls is itself a relative entropy, resolved by the chain rule into per-step conditional divergences. The discarded slack has an exact form in each of three geometries: a Gibbs tilt for the Azuma-Hoeffding and PAC-Bayes bounds, the crossing itself for Ville's and for pooled tests, and a dominating certificate for the $L^p$ maximal bound. That certificate's optional-stopping deficit resolves per step into Bregman divergences of the running maximum. On a path-time space, the same identity gains one factor that prices anticipation: an arbitrary random time carries an e-process ``peeking penalty.'' The partition function can be read as a coalescent--a prefix-sharing probability of independent copies--and geometric mixtures of test martingales gain a pooling benefit for multi-model safe testing.", "5": "2026-08-22T19:13:08.028875"}
{"0": 62, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20326v1", "3": "$TCP_\u03b1$: Margin-Controlled Confidence estimation for reliable Music Information Retrieval", "4": "$TCP_\u03b1$: Margin-Controlled Confidence estimation for reliable Music Information Retrieval. Authors: Parampreet Singh, Anushka Singh, Sumit Kumar, Vipul Arora. Categories: eess.AS cs.LG. Deep neural networks are often overconfident, assigning high confidence even to incorrect predictions. Consequently, users lack a reliable signal for deciding when a prediction can be trusted. Post-hoc confidence estimation addresses this by training a lightweight auxiliary head over a frozen classifier. Existing targets, however, suffer from inherent ambiguity: they assign overlapping confidence values to correct and incorrect predictions, while errors near the decision boundary receive confidence scores indistinguishable from correct predictions. In this work, we propose $TCP_\u03b1$, a novel confidence target that resolves these limitations by introducing a margin-controlled penalty for misclassified samples. We prove that $TCP_\u03b1$ guarantees complete separation between the target values of correct and incorrect predictions, with a separation margin that is independent of the number of classes and increases monotonically with the penalty parameter. Since accurate classifiers naturally produce very few errors, learning these targets results in a severely imbalanced regression problem. We therefore present a systematic study of training strategies for learning under this imbalance and identify an effective training configuration through extensive ablation studies. We evaluate the proposed approach on r\u0101ga identification, investigate its robustness under domain shift, and further validate it on frame-wise ornamentation detection without modifying the selected configuration. Across all settings, $TCP_\u03b1$ consistently outperforms existing confidence targets for failure prediction. Rejecting only the least-confident 8\\% of predictions improves the base model's macro-F1 from 0.89 to 0.98, while fine-tuning the confidence head with only 5\\% labeled samples from a new corpus effectively restores performance under domain shift.", "5": "2026-08-22T19:13:08.032739"}
{"0": 63, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20322v1", "3": "A comparison between ceiling-mounted FMCW, IR-UWB and Wi-Fi radar for in-bedroom human activity monitoring and sleep interruption detection", "4": "A comparison between ceiling-mounted FMCW, IR-UWB and Wi-Fi radar for in-bedroom human activity monitoring and sleep interruption detection. Authors: Anton Lambrecht, Reda El Hail, Xianjun Jiao, Pieter Crombez, Dominique Schreurs. Categories: cs.LG. Despite their growing importance for contact-free radio frequency (RF) based healthcare monitoring, different radio technologies such as frequency-modulated continuous wave (FMCW) radar, impulse radio ultra-wideband (IR-UWB), and Wi-Fi sensing are rarely compared under identical deployment conditions, as existing studies typically differ in hardware, datasets, and evaluation methodologies. In addition, the performance of ceiling-mounted radars, despite their practical deployment and cost advantages in healthcare environments, remain underexplored. Therefore, this paper presents a controlled comparison and analysis of ceiling-mounted FMCW, IR-UWB, and Wi-Fi sensing using synchronized recordings from 20 participants across six room layouts. All technologies are evaluated with the same convolutional neural network (CNN) on both a fine-grained 10-class human activity recognition (HAR) task and a coarse 4-class sleep monitoring task. IR-UWB achieves the highest cross-subject activity recognition performance (89.0% macro F1), while FMCW generalizes best to unseen room layouts (83.8% macro F1). For sleep monitoring, all technologies exceed 92% macro F1 in unseen environments. The results reveal a fundamental trade-off between recognition performance and environmental robustness, which can be explained through differences in range resolution, antenna diversity, Doppler resolution, and spatial information retention. These findings provide practical guidelines for the design of healthcare-oriented RF sensing systems.", "5": "2026-08-22T19:13:08.039060"}
{"0": 65, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20315v1", "3": "Explainable Transformer Models for Clinical Prediction Tasks on Structured Electronic Health Records", "4": "Explainable Transformer Models for Clinical Prediction Tasks on Structured Electronic Health Records. Authors: Jun Ni Du, Lukas Adamek, Maxim Kryukov, Flavio Dormont, Ziv Bar-Joseph. Categories: cs.LG. Predictive models over structured electronic health records (EHRs) remain central to machine learning for healthcare, but few have jointly emphasized quantitative laboratory information and interpretability with respect to input medical events. We present BERT-LER, a BERT-style model for coded EHR timelines pretrained and fine-tuned from a de-identified EHR dataset of 75 million patients, that encodes laboratory test results as discrete tokens while retaining graded information through percentile-based binning, paired with Integrated Gradients for token-level attributions grounded in the input EHR sequence. We benchmark our approach on the public EHRShot benchmark suite and on an asthma severity progression study based on real-world data. This addresses a methodological gap in EHR foundation-style modeling by unifying laboratory value representation and explainability in a single framework, while assessing whether both predictive performance and explanations generalize beyond standard clinical prediction tasks. Across EHRShot and asthma tasks, BERT-LER achieves predictive performance that is competitive with, and on laboratory-related tasks often exceeds, publicly available benchmark models, and provides attributions that align with clinically known risk factors. Our architecture and explainability approach can be applied to many therapeutic areas and prediction tasks using language models trained on structured EHRs.", "5": "2026-08-22T19:13:08.050313"}
{"0": 66, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20295v1", "3": "Physical-Support Confidence Sets for Highly Coherent Dictionaries", "4": "Physical-Support Confidence Sets for Highly Coherent Dictionaries. Authors: Guan-Ju Peng. Categories: cs.LG eess.SP math.ST. Sparse pursuit after dictionary learning can yield a precise atom support even when its physical interpretation is not justified by the calibration data, especially for highly coherent dictionaries where alternative calibration-compatible dictionaries may assign different physical meanings to the same selected support. We develop resolution-aware physical-support inference that jointly accounts for uncertainty in the learned dictionary and in the representation of a deployment signal. Our cross-dictionary confidence correspondence retains calibration-compatible dictionaries and deployment-compatible sparse representations, then projects the surviving explanations onto physical-support space. For local coherent-atom classes with separation scale s, once the deployment data resolve the coherent-block explanation and its atom support, the minimax physical resolution from N calibration signals satisfies $\u03b4_{\\mathrm{opt}}(N,s)\\asymp\\min\\{s,\\frac{1}{\\sqrt{N}s^2}\\}$, with relative resolution governed by the orientation-information scale $Ns^6$. Deployment replication improves physical localization only when orientation changes cannot be absorbed by adjusting the active coefficients. For computation, we introduce active endpoint bracketing (AEB), an adaptive finite-bank procedure that evaluates only candidates that can still affect the physical report and otherwise safely coarsens or abstains. Finite-bank experiments, including a four-region synthetic application, show that a point-valued plug-in selector can be physically overprecise, whereas AEB avoids unsupported refinement with fewer candidate evaluations.", "5": "2026-08-22T19:13:08.056174"}
{"0": 67, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20285v1", "3": "Dynamic Structural Causal Modeling for Sleep", "4": "Dynamic Structural Causal Modeling for Sleep. Authors: Ranveer Singh, Saurabh Mathur, Pranuthi Tenali, Arun Badi, Sriraam Natarajan. Categories: cs.LG. The causal dynamics of sleep-disordered breathing are complex and vary across patient populations, hindering the development of targeted interventions. We learn dynamic causal graphs of sleep-disordered breathing from Home Sleep Apnea Test (HSAT) recordings, revealing systematic differences in causal structure across sex and age subcohorts. We do so using the PCMCI+ algorithm on windowed fractional variables derived from 105 HSAT recordings, exploiting domain knowledge via edge blacklisting and employing bootstrap aggregation to address small subcohort sizes. The learned graphs show that temporal self-dependencies and the apnea-desaturation relationship persist across all cohorts, while other relationships vary substantially.", "5": "2026-08-22T19:13:08.061659"}
{"0": 68, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20280v1", "3": "Which Eviction Policy Should an LLM Cache Use? A Systematic Study Across Workloads, Capacities, and Encoders", "4": "Which Eviction Policy Should an LLM Cache Use? A Systematic Study Across Workloads, Capacities, and Encoders. Authors: Yash Kulkarni, Shubham Harkare, Arvind Suresh Yogesh Babu. Categories: cs.DB cs.LG. Semantic caches reuse an LLM response when the incoming query embedding lies near a cached query, but proposed eviction policies have rarely been compared under one protocol. Using CLEVER, we evaluate FIFO, LRU, LFU, ARC, GDSF, a single-pass streaming adaptation of SISO, and a semantic-redundancy policy across three ordered, deduplicated query corpora, three cache capacities, and two encoders. No evaluated policy improves on LFU by more than 0.041 percentage points in any of the eighteen settings. Replacement is not irrelevant: FIFO and streaming SISO trail LFU by as much as 8.67 and 8.55 points, respectively, at tight capacity.   We explain the missing upside with a conditional packing result. Under exact lookup and insert-on-miss, a newly inserted entry cannot have a resident neighbor within the hit radius, so a geometry-aware eviction rule receives little new redundancy signal. A separate audit exposes a larger problem with the evaluated operating point. At MiniLM's median nearest-neighbor threshold, only 2.1-3.9% of sampled LMSYS and QQP hits are judged answer-substitutable, reducing raw hit rates of 51-60% to quality-adjusted rates of 1.1-2.2%. The cross-encoder study further shows that thresholds do not transfer between embedding models. LFU is the strongest simple default in this protocol; deployment decisions should first establish answer validity and then test sub-point policy differences with exact search.", "5": "2026-08-22T19:13:08.066989"}
{"0": 69, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20258v1", "3": "DICS: Data-Informed Centroid Splitting for Decision Tree Classifiers", "4": "DICS: Data-Informed Centroid Splitting for Decision Tree Classifiers. Authors: MD Saifur Rahman Mazumder, Feng Yu. Categories: cs.LG stat.ML. Decision tree-based models are widely used in machine learning due to their interpretability and strong empirical performance. However, training decision trees can be computationally expensive, particularly for large and high-dimensional datasets, largely due to the exhaustive search over candidate splits at each node. To improve computational efficiency, we propose Data-Informed Centroid Splitting (DICS), a clustering-based framework that constructs a compact and informative set of candidate splits using data-driven priors. By incorporating class-aware structure, DICS significantly reduces the split search space for classification tasks while preserving predictive performance. We further provide theoretical analysis showing that under the stated assumptions, DICS does not degrade the performance of classification trees compared to exhaustive split search. DICS can be incorporated into classification trees, random forests, and gradient-boosting models. Extensive experiments demonstrate that DICS achieves comparable accuracy while substantially reducing training time across synthetic and benchmark datasets, highlighting the benefit of integrating data-informed priors into split selection for scalable classification tree learning.", "5": "2026-08-22T19:13:08.073723"}
{"0": 70, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20255v1", "3": "Transfer Learning in Nonparametric Regression with Deep ReLU Networks", "4": "Transfer Learning in Nonparametric Regression with Deep ReLU Networks. Authors: Junpeng Ren, Carlos Misael Madrid Padilla, Yanzhen Chen, Oscar Hernan Madrid Padilla. Categories: stat.ML cs.LG stat.ME. This paper develops a general transfer learning framework for nonparametric regression with data consisting of multiple groups. Under the assumption that groups share a common structure along with group-specific deviations in additive form, the proposed method employs a two-stage offset learning procedure: the first stage pools data from all groups to estimate an overall mean function, and the second stage estimates offsets for each group, yielding final group-level estimators through additive combination. Upper bounds on the $\\mathcal L_2$ error are established for the proposed framework, covering a broad class of nonparametric estimators under mild complexity and noise conditions. When instantiated with deep ReLU networks, explicit convergence rates are derived under hierarchical composition models, demonstrating the ability to overcome the curse of dimensionality. Conditions that enable positive transfer with faster rates are considered, including learning with simpler functions and data augmentation through pooling samples across groups. Various simulations and real-data experiments further validate the effectiveness of the proposed method.", "5": "2026-08-22T19:13:08.078097"}
{"0": 71, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20222v1", "3": "Gravitational-wave parameter estimation with machine-learning generated surrogate waveforms", "4": "Gravitational-wave parameter estimation with machine-learning generated surrogate waveforms. Authors: Suyog Garg, Kipp Cannon. Categories: gr-qc astro-ph.IM cs.LG. The worldwide network of gravitational-wave detectors have detected more than 350 binary coalescence events till date. Future third-generation detectors, like Einstein telescope, are expected to detect orders-of-magnitude more signals from sources with more complicated characteristics, including eccentric orbits and high-mass ratio binaries. It is well-established that the computational cost of parameter estimation for signals from these kinds of sources will be extremely high. In particular, the process could be sped-up if generating theoretical waveform predictions, used for likelihood calculation becomes faster. Recently, various machine-learning techniques has been proposed to this end. In this work, we propose a two-stage deterministic conditional-autoencoder model for generating four-parameter SEOBNRv4 waveforms. The first-stage of the model generates amplitude and phase series of the waveform, while the second-stage calibrates the residual error in the predictions. Our model achieves a median mismatch of around $10^{-2}$ with the target polarization waveforms, while the calibrated amplitude/phase series achieve $10^{-6}$ level cosine distance error. We then propose a waveform conditioning step to enable use of these surrogate waveforms for downstream parameter estimation tasks. Finally, we perform extensive parameter estimation tests, with ML and EOB waveform injections and try to recover posterior estimates for the source parameters. We find that when ML waveforms are used to recover EOB target parameter estimates, the inferred posterior have some systematic bias. This inherent bias can be estimated and corrected for, and then importance reweighting of posterior samples can enable use of low-accuracy surrogate waveforms at low SNRs.", "5": "2026-08-22T19:13:08.085650"}
{"0": 74, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20186v1", "3": "Decoding silent reading from non-invasive EEG", "4": "Decoding silent reading from non-invasive EEG. Authors: Ingo Marquardt, Anthilia Alchanat, Priyanka Jain. Categories: cs.LG q-bio.NC. Non-invasive decoding of inner speech faces a fundamental data problem: a corpus pairing brain activity with a person's spontaneous inner monologue cannot be collected, and the available proxy paradigms (cued repetitive and retrospectively reported generative inner speech) are slow to acquire, poorly time-locked, and subject compliance is unverifiable. We therefore treat silent reading as a scalable proxy task and ask how much lexical and semantic information a contrastive decoder can extract from it. We report an open-vocabulary analysis of approximately 240,000 word presentations recorded from a single densely-sampled participant across 393 runs (ca. 49 h) of 19-channel dry-electrode EEG. Words from continuous narrative text were presented in rapid serial visual presentation, with typography randomised on every trial to partially decorrelate word identity from low-level visual form. A convolutional EEG encoder, optionally followed by a causal transformer, was trained with a CLIP-style contrastive objective to align short EEG windows with hidden-state embeddings of the presented word taken from a large language model. Decoding, evaluated as word-grouped top-10 retrieval against permutation baselines, was reliably above chance, extended to mid-frequency and rare words, and scaled log-linearly with training-data volume with no sign of saturation. Removing occipital and posterior-temporal electrodes reduced the word-level gain by roughly one third but left context tracking unchanged. Control analyses separate word-level decoding from narrative context tracking and from a non-neural positional prior introduced by the transformer's positional embedding. These results establish that open-vocabulary word-level information is recoverable from EEG during silent reading, and that decoding is data-limited rather than saturated.", "5": "2026-08-22T19:13:08.098331"}
{"0": 75, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20183v1", "3": "Exact Algebraic Computation of Learning Coefficients for Two-Dimensional Singular Models", "4": "Exact Algebraic Computation of Learning Coefficients for Two-Dimensional Singular Models. Authors: Gr\u00e9goire Sergeant-Perthuis, Elias Tsigaridas, Jules Tsukahara. Categories: cs.LG cs.SC math.AG stat.ML. Classical information criteria such as the Bayesian Information Criterion (BIC) rely on regularity assumptions that break down for singular models, leading to incorrect model selection in settings such as deep learning. The Widely Applicable Bayesian Information Criterion (WBIC) relies on local learning coefficients $\u03bb$, which in the analytic case coincides with local Real Log Canonical Thresholds (RLCT) of the Kullback-Leibler divergence of the model, to capture correct marginal likelihood asymptotics. Exact computation of the learning coefficients has been limited to special cases, and only sampling-based estimation methods are generally applicable. We present the first deterministic algorithm that computes local RLCTs exactly for any two-dimensional model whose Kullback-Leibler distance is contact equivalent to a polynomial, derive a bound on its complexity, and demonstrate its effectiveness for a broad class of models, with applications including polynomial neural networks. Beyond providing ground truth to calibrate sampling-based estimators, exact computation reveals algebraic structure in learning coefficients that sampling cannot and out-speeds it in the shallow regime.", "5": "2026-08-22T19:13:08.104563"}
{"0": 77, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20172v1", "3": "Ask Self, Ask Others: Relation Is All You Need", "4": "Ask Self, Ask Others: Relation Is All You Need. Authors: Yuting Ge, Pengju Yang, Mingkai Nie. Categories: cs.LG. Attention directly derives normalized information flow from pairwise scores. We introduce Relation, an alternative token-mixing primitive that first organizes pairwise evidence into explicit Self and Exchange relations and derives information flow afterward. This relational organization gives rise to Full Relation, FlashRelation, Linear Relation, Hybrid Relation, and a KV-style Relation Cache. Across matched decoder-only models at approximately 10M, 30M, and 100M parameters, Full Relation achieves lower final validation NLL than MHA at all three scales. In a fixed-context reference benchmark, FlashRelation is 3.60-4.41x faster than the materialized Full Relation implementation. Across scale-matched production workloads, it reaches 76.4-84.9% of PyTorch FlashAttention throughput while executing the Full Relation operator. Hybrid Relation uses 75% Linear Relation layers and achieves strong language-modeling quality. These results support a relation-first view of token mixing: ask Self, ask Others, then let Flow follow Relation.", "5": "2026-08-22T19:13:08.114755"}
{"0": 79, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20134v1", "3": "Feature Evolution and Migration during Vision Transformer Training", "4": "Feature Evolution and Migration during Vision Transformer Training. Authors: Joonas J\u00e4rve, Halil Ibrahim Aysel, Tarun Khajuria, Meelis Kull. Categories: cs.CV cs.LG. We present a novel view on feature evolution in Vision Transformers (ViTs) by visualizing the training process over two dimensions -- network depth (layer) and training time (epochs). We employ Sparse Autoencoders (SAEs) to extract candidate sparse features from CLS-token representations and compare their activation profiles across epoch--layer pairs. This allows us to study feature-level dynamics that are not directly visible from representation-level similarity measures. Furthermore, we demonstrate how this framework of feature evolution allows us to describe feature migration, the change in the layer where a feature is most detectable during training. Our experiments show that migration is concentrated early in training, occurs more often toward earlier layers than toward deeper layers, and declines as feature organization stabilizes. We further find that deeper layers stabilize earlier and more strongly than shallow layers. The results show that our approach can be employed as a tool for understanding how ViTs learn and evolve.", "5": "2026-08-22T19:13:08.125143"}
{"0": 80, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20123v1", "3": "Discrete Diffusion Inference-Time Control with Nested Sequential Monte Carlo", "4": "Discrete Diffusion Inference-Time Control with Nested Sequential Monte Carlo. Authors: Lohithsai Yadala Chanchu, Hany Abdulsamad, Christian A. Naesseth. Categories: stat.ML cs.LG. We study inference-time control for text generation in discrete diffusion language models, where the goal is to steer sampling toward sequence-level rewards without retraining. Prior work in this domain has focused on particle-based methods such as best-of-$n$ sampling and bootstrap sequential Monte Carlo, which may suffer from overoptimism and weight degeneracy, respectively. We address these limitations using \\emph{nested} sequential Monte Carlo methods. We formulate nested SMC (NSMC) and fully-adapted nested SMC (FA-NSMC) for Feynman--Kac steering, identifying and correcting errors in prior formulations that lead to biased final estimates. We evaluate these methods on toxicity and fluency steering tasks, showing that NSMC and FA-NSMC consistently outperform best-of-$n$ and bootstrap SMC.", "5": "2026-08-22T19:13:08.128982"}
{"0": 81, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20117v1", "3": "SAE-Xplainers: Rule-Based Feature Interpretation for Extreme Earth Events", "4": "SAE-Xplainers: Rule-Based Feature Interpretation for Extreme Earth Events. Authors: Hugo Porta, Emanuele Dalsasso, Chang Xu, Theo Gnassounou, Devis Tuia. Categories: cs.LG. The emergence of large-scale Weather and Climate (W&C) datasets offers new opportunities for modeling extreme Earth events (ExEE) and their impacts using deep learning. However, their adoption in operational settings remains limited by the lack of models' interpretability. While for conventional text and image modalities, tools such as Sparse Autoencoders (SAEs) have proven effective for extracting human-understandable concepts, their use for the analysis of ExEE remains challenging due to the nature of W&C data. To address this, we introduce (i) a geographic location-based modulation of the inputs of SAE to capture the local semantic meaning of environmental patterns, and (ii) an ensemble of rule-based SAE-Xplainers to interpret the resulting high-dimensional features derived from complex, multi-modal environmental predictors. We evaluate our method on three ExEE types: the prediction of fires, and the detection of tropical cyclones and atmospheric rivers. We show that SAE input modulation improves both reconstruction performance and feature utilization, and that our SAE-Xplainers enable faithful interpretation of complex climatic patterns by unfolding them into human-understandable rules that are consistent with the scientific literature, while also supporting the identification of feature absorption.", "5": "2026-08-22T19:13:08.133323"}
{"0": 84, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20065v1", "3": "Orthogonal JEPA: Factorized Predictive States for Latent World Models", "4": "Orthogonal JEPA: Factorized Predictive States for Latent World Models. Authors: Taoyong Cui, Pheng Ann Heng, Wanli Ouyang. Categories: cs.LG. World models construct latent states that support prediction, planning, and reasoning about an underlying system. Joint-embedding predictive architectures (JEPAs) offer a direct way to learn such states by predicting targets in representation space instead of reconstructing every detail of the observation. Standard JEPAs, however, organize all predictable content through one target embedding and one prediction pathway. In complex systems, this monolithic state can allocate redundant capacity to dominant signals while providing weak or conflicting gradients to less dominant predictive structure. We introduce \\method, a latent world-modeling framework based on orthogonal predictive factorization. Learned basis matrices analyze each target state into multiple components, and a dedicated prediction branch estimates each component from a shared context representation. Predictive regression preserves the factor magnitudes required for state synthesis, an orthogonality objective discourages repeated directions, factor-activity regularization maintains variation in projected targets, and online variance regularization discourages coordinate-wise encoder collapse. Predicted components are synthesized into a complete latent state that can be used by a readout, decoder, planner, or autoregressive rollout. The same predictive-state mechanism applies when the target is temporally future, spatially hidden, or another partial observation of the same system. Experiments on controlled vision, single-cell transcriptomics, longitudinal health records, continuous control, and molecular dynamics evaluate representation quality, forecasting, planning, and long-horizon stability.", "5": "2026-08-22T19:13:08.151650"}
{"0": 86, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20054v1", "3": "What You Can't See Is What You Learn: Restricted Evidence Visibility Favors Compositional Generalization in Shared-Genome Language-Model Societies", "4": "What You Can't See Is What You Learn: Restricted Evidence Visibility Favors Compositional Generalization in Shared-Genome Language-Model Societies. Authors: Narcis Marincat. Categories: cs.AI cs.LG cs.MA. Multi-module systems often expose every module to the full input. We test whether restricting evidence visibility changes which solutions gradient-based training discovers. Four-cell societies share one frozen pretrained language model and one low-rank adapter, communicating only through two model-width continuous vectors in a fixed relay. On a prospectively sealed natural-language function-composition task, we train ten matched restricted/global pairs sharing initialization bytes, training order, token layout, parameters, and computation; only the attention mask differs. Restricted societies outperform their globally visible twins by at least 20 points at both depths in 9 of 10 pairs, with median paired advantages of 0.7648 and 0.6050. Cutting communication reduces every restricted society to chance, and the depth-three advantage remains 0.558 on programs whose composite function never appeared in training. Across six audited restricted societies, same-value packet transplants preserve behavior at 0.94-1.00 across all tested interfaces; destructive interventions collapse performance; and counterfactual packets redirect outputs toward the mathematically predicted answer. The sole high-performing global model also requires communication, but its same-value packets are not interchangeable across episodes. Restricted visibility is thus not necessary for composition; under this protocol it substantially increases the probability of a generalizing relay and favors a reusable, value-indexed interface. The complete preregistered battery nevertheless formally fails because restricted-arm median depth-three accuracy is 0.6988, below the 0.70 floor. An earlier qualification cohort likewise yielded 0/10 complete passes: one model met every task-performance gate, but all ten failed ordinary-language preservation, confining the system to explicitly task-gated use.", "5": "2026-08-22T19:13:08.161121"}
{"0": 87, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20052v1", "3": "DecoVAE: a Lightweight Interpretable Trend-Seasonal VAE Framework for Efficient Probabilistic Time Series Forecasting", "4": "DecoVAE: a Lightweight Interpretable Trend-Seasonal VAE Framework for Efficient Probabilistic Time Series Forecasting. Authors: Alexander Marusov, Dmitry Anikin, Alexey Zaytsev. Categories: cs.LG. Probabilistic time series forecasting remains challenging, largely because modeling distinct trend and seasonal dynamics requires specialized approaches. Existing methods often fail to capture the unique inner properties of these components, lack interpretability, or suffer from heavy memory and runtime overhead. To address these limitations, we propose DecoVAE, a lightweight interpretable trend-seasonal VAE framework that explicitly decomposes time series into trend and seasonal components by applying domain-specific inductive biases. The trend stream enforces structural smoothness using a differential regularizer on the latent trajectory, analogous to the Hodrick-Prescott filter. Concurrently, the seasonal stream operates in the frequency domain via a complex Gaussian VAE, natively capturing the amplitude and phase of periodic patterns. Extensive evaluations across seven real-world benchmarks show that DecoVAE consistently outperforms strong baselines. It achieves reductions of up to 14.96\\% in CRPS and 23.30\\% in NMAE for short-term forecasting, and up to 52.68\\% and 26.51\\% for long-term horizons. Crucially, DecoVAE yields these accuracy gains while remaining highly efficient, reducing model weight by up to 93\\% and accelerating speed by up to 74\\% compared to the second-best method.", "5": "2026-08-22T19:13:08.167180"}
{"0": 89, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20044v1", "3": "End-to-end Early Classification of Time Series in Non-Stationary Environments", "4": "End-to-end Early Classification of Time Series in Non-Stationary Environments. Authors: Aur\u00e9lien Renault, Alexis Bondu, Antoine Cornu\u00e9jols, Vincent Lemaire. Categories: cs.LG. Early Classification of Time Series (ECTS) requires making accurate decisions as early as possible in inherently online and evolving environments. Yet, most existing methods assume stationarity and rely on separable designs, where classification and triggering are optimized independently, an assumption that fundamentally limits their adaptability under drift. In this work, we challenge this paradigm and study ECTS under non-stationary conditions. We provide the first systematic comparison between separable and end-to-end approaches across controlled drifting scenarios. Building on Reinforcement Learning, we introduce DQeND, a unified architecture that jointly learns representation, classification, and triggering decisions, while remaining directly comparable to state-of-the-art separable baselines. Across a wide range of drifts, DQeND demonstrates strong robustness across various non-stationary scenarios, consistently outperforming separable baselines. An ablation study further highlights that jointly updating representation and decision modules is critical to these gains. Overall, our results indicate that end-to-end learning can offer improved adaptation capabilities for ECTS in dynamic environments, and motivate further investigation of alternatives to separable designs.", "5": "2026-08-22T19:13:08.178636"}
{"0": 90, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20038v1", "3": "An Inclusive and Lightweight Approach to Federated Continual Learning for Cultural Heritage", "4": "An Inclusive and Lightweight Approach to Federated Continual Learning for Cultural Heritage. Authors: Ioannis Theologitis, Debin Meng, Stylianos Eleftheriadis, Vasileios Lolis, Konstantinos Votis. Categories: cs.LG cs.AI cs.CV. Artificial intelligence can support cultural heritage and digital humanities through large-scale retrieval and analysis of digitized collections. However, cultural heritage data are often distributed across institutions, constrained by ownership and access restrictions, and continuously evolving over time. Federated Continual Learning (FCL) is well suited to this setting, as it enables models to learn from distributed and sequential data without sharing raw collections. In this paper, we propose FedCurv-DR, a lightweight, regularisation-based FCL strategy. The method accumulates parameter-importance estimates across clients and experiences to protect learned knowledge, while updating them only at fixed intervals to minimize communication and computation overhead. We evaluate FedCurv-DR in a continual learning scenario using the WikiArt image dataset for genre classification with evolving styles, reporting performance, energy, and fairness metrics. Our results show that FedCurv- DR reduces forgetting and balances performance, fairness, and energy efficiency for sustainable AI in cultural heritage.", "5": "2026-08-22T19:13:08.184636"}
{"0": 92, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20160v1", "3": "Chameleon: Robust Defense Against Tor Website Fingerprinting via Many-to-Many Traffic Morphing", "4": "Chameleon: Robust Defense Against Tor Website Fingerprinting via Many-to-Many Traffic Morphing. Authors: Yuwen Cui, Kai Wei, Kehan Shen, Ning Wang, Zhuo Lu. Categories: cs.CR cs.CY. Website fingerprinting (WF) attacks can infer users' browsing activities from encrypted Tor traffic by exploiting side-channel features. Although many WF defenses have been proposed, we find that most existing defenses create learnable web trace mapping features. We further show that robustness against adversarial training does not necessarily imply robustness against defense-aware autoencoder (DAAE)-based attacks.   To address these limitations, we present Chameleon, a robust WF defense based on many-to-many randomized traffic morphing. Chameleon selects morphing candidates with high intra-class diversity and low inter-class disparity. Chameleon randomly maps each webpage trace to multiple candidates, and allows different webpages to share morphing targets, thereby increasing adversarial uncertainty. For practical Tor deployment, Chameleon introduces a radix-trie-based synchronization mechanism that enables pluggable transport (PT) endpoints to identify consistent morphing traces using packet-direction prefixes, together with trace mutation and normalized prefix matching to reduce overhead. We evaluate Chameleon against six state-of-the-art defenses and five WF attacks on three public datasets in closed- and open-world settings. Compared with Adaptive Tamaraw, Chameleon reduces adversarial-training-based attack accuracy by up to 36.74% while reducing bandwidth and time overhead by 34.12% and 60.38%, respectively. Under DAAE-based RF attacks on GTT23, Chameleon limits attack performance to 35.19% F1-score while Adaptive Tamaraw only limits it to 88.22% F1-score. In the real-world PT bridge evaluation, Chameleon substantially reduces the effectiveness of strong WF attacks while incurring only 16.25% time overhead.", "5": "2026-08-22T19:13:11.830622"}
{"0": 93, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20145v1", "3": "Trustworthy mobile edge caching: a blockchain approach to mitigate malicious nodes and incentivize cache sharing", "4": "Trustworthy mobile edge caching: a blockchain approach to mitigate malicious nodes and incentivize cache sharing. Authors: Motahare Ebrahimi, Nastooh Taheri Javan, Seyedakbar Mostafavi, Fatemeh Pakzaban. Categories: cs.CR cs.CY. As mobile network traffic continues to grow, content caching on edge servers is critical for reducing latency. However, challenges such as malicious edge servers that may delete or manipulate cached content, along with the limited capacity of these servers, need to be addressed. To overcome the capacity limitations, helper mobile nodes can contribute their cache resources. However, due to their selfish behavior, an incentive mechanism is necessary to encourage resource sharing. Additionally, these helper nodes can also be malicious. This paper proposes a blockchain-based trust management mechanism that addresses these challenges by accurately identifying trustworthy edge servers and mobile nodes. The proposed mechanism calculates both direct and indirect trust using smart contracts, ensuring that malicious nodes are effectively filtered out. Trustworthiness is determined based on mobile node satisfaction with the quality of service, and trust data is securely stored on the blockchain. To combat node selfishness, a reward mechanism is introduced to incentivize cache sharing. Furthermore, a blockchain-based authentication mechanism protects against node impersonation. Our approach optimizes trust, cache capacity, and cost efficiency while considering mobile node mobility, energy consumption, and computational power constraints during the consensus process. Simulation results show that the proposed method can accurately distinguish between honest and malicious servers, even with a 10% noise in data.", "5": "2026-08-22T19:13:11.837265"}
{"0": 94, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20118v1", "3": "Privacy-Preserving Detection of Rare Disease-Associated Cell Subsets via Secure Multi-Party Computation", "4": "Privacy-Preserving Detection of Rare Disease-Associated Cell Subsets via Secure Multi-Party Computation. Authors: \u015e. Selcan Magara, Esther Havemann, Debora Jutz, Ali Burak \u00dcnal, Mete Akg\u00fcn. Categories: cs.CR. The detection of rare disease-associated cell subsets from high-dimensional single-cell measurements is critical for understanding diseases such as leukaemia and viral infections. CellCnn, a convolutional neural network (CNN) designed for this task, has demonstrated the ability to identify phenotype-associated cell populations at frequencies as low as 0.01\\%. Training such models reliably requires patient cohorts that are larger and more diverse than any single institution can typically assemble, and the underlying single-cell data is too sensitive to share across institutional boundaries under existing privacy regulations. We propose a secure multi-party computation (MPC) framework that enables the training and inference of CellCnn entirely on secret-shared data. This ensures that neither the participants nor the computing servers ever observe raw patient data or intermediate values. Evaluated on benchmark single-cell datasets for cytomegalovirus infection (CMV) and acute myeloid leukaemia (AML), our implementation preserves accuracy close to its plaintext counterpart while outperforming the prior privacy-preserving baseline. In contrast to earlier privacy-preserving approaches that removed components such as ReLU activations and bias terms, our method retains these key parts of the CellCnn architecture and supports accurate analysis without exposing raw patient data.", "5": "2026-08-22T19:13:11.842601"}
{"0": 95, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20108v1", "3": "A Meta-Study on Replication Papers in Usable Security & Privacy", "4": "A Meta-Study on Replication Papers in Usable Security & Privacy. Authors: Christian Mack, Benjamin Berens, Hanna Algedri, Tobias Hilt, Daniela Reimer. Categories: cs.CR. The field of usable security and privacy research is a young and expanding field, which is still developing standards for its research, e.g. regarding replications. We used a mixed-method approach, in order to get a better understanding of the current state of replications in the field of usable security and privacy: (1) we examine the Call for Papers of 13 venues spanning security, privacy, and human-computer interaction; (2) we conduct a systematic search for papers reporting replicated user studies published across these venues between 2016 and 2025, yielding 24 relevant publications; (3) we categorized these 24 papers employing the replication taxonomy proposed by Olszewski et al. (2025); (4) we distributed a survey to the authors of these papers to understand their motivations for conducting replications. Our analysis reveals four key insights: (A) Calls for Papers would benefit from clearer guidelines for authors and reviewers regarding replication work; (B) determining what modifications were made relative to the original study proves difficult when reading replication papers; (C) strict exact replications do not exist in our sample. Approximately two-thirds of the 24 studies altered multiple aspects of the original work; (D) temporal and contextual changes affecting results emerged as one of the most frequently cited motivations for replication. Based on these findings, we offer practical recommendations for venues, researchers, and peer reviewers to strengthen replication practices in usable security and privacy research.", "5": "2026-08-22T19:13:11.848552"}
{"0": 96, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20097v1", "3": "TrustRAG: Blockchain-Enhanced RAG via Committee-Based Credibility Scoring", "4": "TrustRAG: Blockchain-Enhanced RAG via Committee-Based Credibility Scoring. Authors: Baixiang Liu, Haotian Che, Yuan Li. Categories: cs.CR cs.DC. Retrieval-Augmented Generation (RAG) lets Large Language Models (LLMs) pull in up-to-date, domain-specific information instead of relying only on what they were trained on. Yet most RAG systems still draw from centralized databases with limited oversight, making it difficult to verify where a document came from, whether it has been tampered with, or whether it should be trusted at all. This is a serious problem in domains where both the timeliness and accuracy of retrieved content are critical, such as healthcare, finance, logistics, and legal case law, where a wrong or manipulated document can directly lead to bad decisions.   We present TrustRAG, a committee-based, blockchain-backed RAG system: before a document is used, it is certified by a committee of domain experts through a zero-knowledge protocol, and the committee's hidden scores are combined via secure multi-party computation into a trust score that any client can verify. These scores, along with the underlying document data, are maintained jointly across chains through hash commitments, so no document or score can be silently altered or dropped, and every ranking can be independently replayed and checked.", "5": "2026-08-22T19:13:11.852903"}
{"0": 97, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20055v1", "3": "EchoCoT: Extracting Hidden Chain-of-Thought from Large Reasoning Models", "4": "EchoCoT: Extracting Hidden Chain-of-Thought from Large Reasoning Models. Authors: Yiting Qu, Ziqing Yang, Chi Cui, Ye Leng, Junjie Chu. Categories: cs.CR cs.AI. Hidden chain-of-thought (CoT) traces, especially those from frontier proprietary large reasoning models (LRMs), are valuable model assets. Yet whether these hidden CoTs can be directly extracted from black-box models remains largely unexplored. In this work, we systematically study whether hidden CoTs can be extracted near-verbatim from black-box LRMs through API interactions. We identify a previously overlooked reasoning replay surface between tool calls and develop EchoCoT, a multi-step attack that iteratively extracts hidden CoTs using API-returned fidelity signals. We further develop an LLM-based optimization framework that automatically searches for an effective universal injection trajectory across various datasets. We evaluate EchoCoT on three open-source and five frontier proprietary LRMs. On open-source LRMs, EchoCoT achieves up to 66.4\\% near-verbatim extraction success, with the extracted trace length within 10\\% of the target and at least 90\\% of tokens exactly matching the target CoT. The same injection trajectory also generalizes to unseen datasets, achieving up to 80\\% extraction success under the same criterion. For tested frontier proprietary LRMs, a substantial fraction of extracted CoTs closely align with provider-reported reasoning lengths and available CoT summaries. EchoCoT can also extract very long CoTs: on Gemini-2.5, it extracts 33,463 tokens from a 32,948-token target. These results establish hidden-CoT extraction as a practical security risk and highlight the need to better protect hidden CoT assets.", "5": "2026-08-22T19:13:11.860629"}
{"0": 99, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19982v1", "3": "COPA: Continual Preference Optimization for Adaptive Prompt Injection Defense", "4": "COPA: Continual Preference Optimization for Adaptive Prompt Injection Defense. Authors: Roshan Sood, Onat Gungor, Tajana Rosing. Categories: cs.CR. LLMs remain vulnerable to prompt injection attacks, where adversarial instructions embedded in user inputs or external content manipulate model behavior and bypass safeguards. Existing defenses are predominantly static, relying on fixed alignment objectives or attack-specific filtering mechanisms that require redesign as new attack strategies emerge. While recent lifelong alignment methods address shifting user preferences, they do not account for adaptive adversaries that continually evolve to exploit weaknesses in previously learned defenses. This limitation is particularly important in real-world deployments, where evolving attack distributions necessitate continual adaptation without sacrificing robustness to previously encountered threats. We present COPA, a continual preference optimization framework that treats prompt-injection defense as a lifelong learning problem. Instead of one-time alignment, COPA incrementally incorporates feedback from newly observed attacks via GRPO-based optimization and uses margin-weighted experience replay to retain defenses against prior attack classes. This enables continuous adaptation to emerging threats while mitigating catastrophic forgetting and preserving general-purpose model capabilities. Across lifelong prompt injection attack streams, COPA reduces attack success rate by up to 6.3x and 4.4x on average compared to state-of-the-art defenses. These results highlight continual preference optimization as an effective paradigm for defending LLMs against adaptive adversaries.", "5": "2026-08-22T19:13:11.869491"}
{"0": 100, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19959v1", "3": "Tracking the Trend in How Speech Synthesizers Deceive People", "4": "Tracking the Trend in How Speech Synthesizers Deceive People. Authors: Milan \u0160alko, Anton Firc, Kamil Malinka, Vojt\u011bch Stan\u011bk, Martin Pere\u0161ini. Categories: cs.CR cs.SD eess.AS. Advances in speech synthesis have made deepfake audio highly realistic. Earlier studies reported 70-80% human detection accuracy, but relied primarily on older synthesizers. We compare human detection for three selected voice synthesis tools released in 2019, 2022, and 2024 with 82 IT professionals, and benchmark humans against six pretrained detectors on the same material. For fully synthetic speech (full spoofs), the F1 score drops from about 90% for RTVC and YourTTS to 48% for ElevenLabs, although listeners were explicitly warned that deepfakes were present. For partial spoofing, where only one sentence of an utterance is altered, strict accuracy falls to 9%, and listeners classify the synthetic sentence as bona fide 77% of the time. Humans and detectors fail in complementary ways, and neither reliably localizes short manipulations. Additionally, listeners increasingly mislabel bona fide speech as fake, eroding trust in unmanipulated audio. These findings show that human perception alone is unreliable for the selected modern and partial-spoof conditions and motivate procedural verification, provenance, watermarking, and segment-level detection.", "5": "2026-08-22T19:13:11.876378"}
{"0": 101, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19938v1", "3": "From Noise to Signal: Improving Security Log Anomaly Detection Using LLMs with Endpoint-Specific Logs", "4": "From Noise to Signal: Improving Security Log Anomaly Detection Using LLMs with Endpoint-Specific Logs. Authors: Christopher Henshaw, Gour Karmakar. Categories: cs.CR cs.LG. Existing approaches to anomalous behaviour log detection, such as Wazuh rely primarily on predefined detection rules, while statistical anomaly detection approaches such as OpenSearch identify deviations from previously observed behavioural patterns. Recent research has investigated LLMs for log anomaly detection because of their ability to interpret semantic and contextual information. However, LLM-based approaches can be affected by prompt construction, noisy log data, and reliance on generic datasets that may lack endpoint-specific authentication behaviours. To address these limitations, this study develops a standardised instruction-based LLM classification framework for detecting anomalous authentication behaviours, including borderline cases. A controlled cybersecurity testbed was developed to generate endpoint-specific authentication data, producing a curated dataset comprising normal, borderline, and anomalous behavioural scenarios. Three instruction-tuned LLMs, Meta Llama 3.1 8B Instruct, Qwen 2.5 7B Instruct, and GPT-OSS 20B, were evaluated against Wazuh rule-based detection and OpenSearch Anomaly Detection using a common ground-truth severity framework. Meta Llama 3.1 8B Instruct achieved the strongest overall end-to-end detection performance, with an accuracy of 89.3%, recall of 88.2%, F1-score of 91.8%, and false negative rate of 11.8%. In comparison, Wazuh achieved an accuracy of 52.0% and false negative rate of 68.6%, while OpenSearch achieved an accuracy of 49.3% and false negative rate of 74.5%. Meta Llama also detected 80% of the borderline anomalous scenarios, compared with 20% for Wazuh and 15% for OpenSearch. Qwen achieved lower overall detection performance than Meta Llama but recorded the lowest average inference latency and 100% structured-response validity. GPT-OSS demonstrated strong classification performance when valid responses were produced.", "5": "2026-08-22T19:13:11.881216"}
{"0": 102, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19937v1", "3": "ShadowPath: Lookup-Private Credential Status Verification over Authenticated State", "4": "ShadowPath: Lookup-Private Credential Status Verification over Authenticated State. Authors: Patrick Herbke, Wolf Rieder, Christian Ren\u00e9 Sechting, Huaning Yang, Sid Lamichhane. Categories: cs.CR. Verifiable credentials let holders present digitally signed claims without requiring the issuer to participate in every presentation. Revocation complicates this privacy model because a verifier must determine whether a credential remains valid. Existing status checks may expose recurring identifiers, registry positions, or request metadata. Such information can serve as stable handles to link separate presentations. ShadowPath moves the credential status lookup to the holder. For each presentation, the holder proves, in zero-knowledge, that the credential has not been revoked under the verifier-selected registry root. The verifier learns the status result but not observable metadata. To the best of our knowledge, we provide the first evaluation of Verkle trees for credential revocation and compare them with sparse Merkle trees to assess their applicability in real world applications. The comparison tests whether reducing path depth with Verkle trees offsets the higher cost of KZG-based authentication. Across 30 desktop trials, median Groth16 proving took 371.6ms with sparse Merkle and 2.11s with Verkle. Verification took 3.70ms and 7.55ms, respectively. Groth16 Verkle proving took about 3s on both primary mobile devices. The results show that shorter authenticated paths do not necessarily yield cheaper zero-knowledge proofs. With fresh session randomness, verifier-visible status data do not reveal whether two presentations use the same credential under the stated assumption of session-value independence. This guarantee excludes issuer-verifier collusion and synchronization traffic.", "5": "2026-08-22T19:13:11.886642"}
{"0": 103, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19924v1", "3": "Securing Filesystems for Confidential Computing", "4": "Securing Filesystems for Confidential Computing. Authors: Dimitra Giantsidi, Antoine Delignat-Lavaud, C\u00e9dric Fournet, Jinnan Guo, Heidi Howard. Categories: cs.CR. Confidential computing protects applications inside Trusted Execution Environments (TEEs), but it leaves storage vulnerable. Even with disk encryption, a malicious cloud provider can roll back, replay, fork, or tamper with disk state, breaking the integrity and freshness guarantees required by stateful applications. Existing solutions either assume trusted storage, incur high overheads, or push integrity logic into applications. We present ShieldFS, a POSIX-compliant filesystem that provides end-to-end integrity and freshness for persistent storage in the confidential-computing threat model without requiring application changes. ShieldFS represents permissible filesystem states using succinct cryptographic commitments, maintained inside TEEs and replicated in a lightweight trusted registry. On-disk data structures, including a write-ahead log and a storage pool, are authenticated using hash chains and an embedded Merkle tree. ShieldFS utilizes transactions and copy-on-write to update persistent filesystem state and commitments atomically. The commitments are verified during reads, ensuring that rollback, replay, and equivocation attacks are detected even when the entire I/O stack is untrusted. We implement the design by extending ZFS, yielding ShieldZFS. Evaluation with standard filesystem benchmarks and real-world workloads shows that ShieldZFS provides strong integrity and freshness guarantees with performance comparable to state-of-the-art filesystems.", "5": "2026-08-22T19:13:11.894743"}
{"0": 104, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19901v1", "3": "MaliciousSkillBench: A Comprehensive Benchmark for Malicious Agent Skill Detection", "4": "MaliciousSkillBench: A Comprehensive Benchmark for Malicious Agent Skill Detection. Authors: Yue Wang, Yi Liu, Gelei Deng, Ying Zhang, Yuekang Li. Categories: cs.CR cs.AI. Agent Skills extend LLM agents with reusable instruction packages that may also include scripts, resources, and service configuration. This creates a direct distribution channel for malicious behavior, yet existing malicious-Skill datasets are fragmented across sources, artifact formats, evidence regimes, and benign coverage; duplicated and structurally related content further complicates direct aggregation and evaluation. We present MaliciousSkillBench, a comprehensive benchmark for malicious Agent Skill detection. We consolidate 13 public sources, 11 of which contribute Core malicious artifacts, and reduce 8,414 raw malicious records to 7,539 normalized-unique identities in 4,588 operational structural families. After conservative cross-label conflict exclusion, the primary benchmark contains 9,740 Skills: 7,505 malicious and 2,235 benign. To characterize its coverage, we harmonize 11 attack categories for 4,983 malicious identities with supported source-native mappings and find substantial differences in threat composition across sources. We then evaluate three learned text detectors and three off-the-shelf Skill scanners. Learned detectors achieve 0.882-0.932 Random Macro-F1 but only 0.653-0.665 under Source-Disjoint evaluation; the strongest word TF-IDF SVM scores 0.932/0.916/0.665 on Random/structural-disjoint/Source-Disjoint while retaining 95.6% malicious recall but producing 62.4% benign FPR on held-out sources. Off-the-shelf scanners occupy different but also unsatisfactory operating regimes, reducing false positives only at the cost of sharply lower malicious recall. Together, these results show that reliable malicious-Skill detection requires both broader cross-source benchmark coverage and evaluation that jointly measures attack detection and benign over-flagging.", "5": "2026-08-22T19:13:11.898404"}
{"0": 105, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19857v1", "3": "Inadvertent Context Leakage in Language Models", "4": "Inadvertent Context Leakage in Language Models. Authors: Jaiden Fairoze, Neal Mangaokar, Kamalika Chaudhuri, Sanjam Garg, Saeed Mahloujifar. Categories: cs.LG cs.CR. For AI agents to be useful beyond simple chat, they must hold sensitive user context such as calendars, credentials, health records, and financial data. We study whether the mere presence of such secrets in a model's context window introduces hidden correlations into the model's benign outputs, allowing reconstruction even when the model correctly refuses direct extraction. We further study whether an adversary can actively engineer prompts that amplify this effect, using the model as a covert carrier to transmit secrets through seemingly innocuous text. In both cases, this limited leakage is exploited using a novel adaptive attack that assumes black-box access to the underlying model.   In controlled experiments across eight proprietary models, we find that 2-digit in-context secrets are reconstructed with near-perfect accuracy and 4-digit secrets at 82\\% exact match, all from outputs the model produces in response to ordinary, non-adversarial requests. We observe that more capable models leak more: stronger instruction-following amplifies sensitivity to in-context secrets, suggesting leakage is a byproduct of capability as opposed to a patchable bug. We show this leakage enables two practical attacks: (1) a trained classifier that infers semantic predicates about user memories (e.g., health conditions, financial events) from routine natural-language outputs, and (2) an RL-trained adversary that extracts full Social Security Numbers from a production-style agent.", "5": "2026-08-22T19:13:11.904551"}
{"0": 106, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19821v1", "3": "Survival of~the~Stealthiest: Evolving Low-Entropy Ransomware via~Genetic Algorithms", "4": "Survival of~the~Stealthiest: Evolving Low-Entropy Ransomware via~Genetic Algorithms. Authors: Efrat Levenberg, Kristina Sviazhina, Ayelet Butman, Pierre Parrend, Harel Berger. Categories: cs.CR. Traditional ransomware deployment often relies on massive encryption procedure, triggering immediate detection by modern defense systems. This work introduces a paradigm shift in cryptographic attacks by framing ransomware execution as a Search-Based Software Engineering (SBSE) optimization problem. This approach addresses the persistence gap observed in modern threats, where attacks aim to remain undercover for hours rather than minutes. Using a Genetic Algorithm (GA), we optimize data encryption under a hard constraint on the statistical deviation from baseline system activity. We demonstrate that our evolved attack patterns can evade behavioral monitors under fingerprinting techniques. Our results suggest that search-based methods provide a powerful framework for generating evasive malware, highlighting an emerging challenge for automated software defense.", "5": "2026-08-22T19:13:11.909192"}
{"0": 107, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19750v1", "3": "TGL-APT: Temporal Graph Learning with Graph Distillation for Efficient APT Investigation", "4": "TGL-APT: Temporal Graph Learning with Graph Distillation for Efficient APT Investigation. Authors: Jing Chen, Ayong Ye, Yuanhuang Liu, Yuexin Zhang. Categories: cs.CR. Advanced Persistent Threat (APT) attacks pose a critical challenge to modern systems, as their stealthy, multi-stage nature renders conventional detection methods ineffective. While provenance graphs provide rich behavioral context for attack investigation, attack-relevant evidence is often sparse and embedded in large volumes of routine system activity, making full-graph learning both computationally expensive and difficult to correlate over long attack sequences. We present TGL-APT, an adaptive investigation framework built on the observation that attack-relevant information is non-uniformly distributed and often mediated by structurally influential or behaviorally distinctive entities, which we characterize as information-bottleneck nodes. TGL-APT combines three complementary components: (1) information-bottleneck-guided graph distillation that suppresses provenance redundancy while bounding structural distortion and preserving causal reachability; (2) adaptive temporal graph learning that continuously refines the core node set as node relevance evolves; and (3) cross-spatiotemporal attack fingerprint alignment that associates fragmented suspicious activities across different entities and time windows. Finally, causal expansion and stage characterization reconstruct coherent attack processes for investigation. Experiments on three DARPA E3 datasets show F1-scores of 95.7%, 90.9%, and 88.9%, while reducing training time, detection latency, and memory usage by approximately 39%, 33%, and 22%, respectively, compared with KAIROS. These results demonstrate that TGL-APT effectively balances detection performance, computational efficiency, and investigation capability for provenance-based APT analysis.", "5": "2026-08-22T19:13:11.914610"}
{"0": 108, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19650v1", "3": "Enhancing Privacy in Federated Learning via Dual Obfuscation of Gradients and Training Images", "4": "Enhancing Privacy in Federated Learning via Dual Obfuscation of Gradients and Training Images. Authors: Yuki Itabashi, Hiroto Sawada, Mare Hirose, Shoko Imaizumi, Hitoshi Kiya. Categories: cs.CR. Federated learning enables collaborative model training while keeping data locally at each client; however, recent studies have shown that training data can be reconstructed from shared model updates. To address this issue, this paper proposes a dual obfuscation method that enhances robustness against image restoration attacks by jointly obfuscating updated information and training images. The proposed method combines a robustness enhancement technique based on random binary weights, which randomly sets a portion of gradient elements to zero, with an image encryption technique. These techniques provide complementary protection by reducing the amount of original gradient information available to an attacker and the visual interpretability of reconstructed images, respectively. Furthermore, the image encryption technique allows independent keys to be used for each client and each image, avoiding explicit key sharing. Experimental results on an image classification task using a Vision Transformer (ViT) show that the proposed method reduces the visual information recovered by Attention Privacy Leakage (APRIL) under the evaluated settings without causing additional degradation in classification performance beyond that caused by image encryption. Although the proposed combination does not provide an absolute security guarantee, the results demonstrate the potential benefit of combining gradient modification and image encryption for privacy-enhanced federated learning.", "5": "2026-08-22T19:13:11.919917"}
{"0": 109, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19534v1", "3": "AEGIS: Attention-Embedding Gradient Isolation Shield - Triple-Channel Gradient Masking for Privacy-Preserving Federated LLM Fine-Tuning", "4": "AEGIS: Attention-Embedding Gradient Isolation Shield - Triple-Channel Gradient Masking for Privacy-Preserving Federated LLM Fine-Tuning. Authors: Ye Tao, Hong Shen, Hui Tian, Xin Wang, Can Wang. Categories: cs.CR. Gradient inversion attacks recover private training text from gradients shared in federated learning, posing a serious threat to collaborative model training. Through our analysis of transformer gradient structure, we identify three channels through which private token information leaks: the attention output projection gradient exposes a low-rank subspace that encodes input embeddings (Channel 1), the embedding gradient's row-norm sparsity directly reveals which tokens are present (Channel 2), and the MLP expansion gradient carries a recoverable subspace signal analogous to Channel 1 (Channel 3). State-of-the-art attacks exploit these channels analytically to achieve near-exact token recovery in seconds. Existing defences address at most one channel and either degrade model utility or leave the remaining structural signals intact.   We introduce AEGIS (Attention-Embedding Gradient Isolation Shield), a lightweight defence that closes all three analytical channels with three backward-path operations requiring no architectural changes: freezing attention projection parameters eliminates Channel 1 by construction, calibrated noise injection into the embedding gradient destroys Channel 2's token-presence signal, and analogous per-block noise injection into the MLP expansion gradient masks Channel 3. The same masked gradient drives both the local optimiser step and the server export, so no clean signal is retained on either side.   Evaluated across 11 models and six datasets, AEGIS reduces token recovery rates to near zero against a range of gradient inversion attacks, both analytical and optimisation-based, while preserving or improving model utility. We provide formal guarantees for Channels 1 and 2 and validate the full defence empirically against adaptive adversaries with complete knowledge of the mechanism.", "5": "2026-08-22T19:13:11.925792"}
{"0": 110, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19462v1", "3": "A Federated Learning Framework for Privacy-Preserving Oral Cancer Screening on Smartphones", "4": "A Federated Learning Framework for Privacy-Preserving Oral Cancer Screening on Smartphones. Authors: Lena D. Swamikannan, Akshay Bhagwan Sonawane, Jay S. Patel, C. S. Mani, Lakshmi Narayana. Categories: cs.CR. Data are the cornerstone of robust AI models. However, in the medical domain, access to reliable data is constrained by regulatory requirements and patient privacy, and clinical oral images are particularly difficult to obtain. Federated learning (FL) mitigates these constraints by enabling collaborative model development across decentralized datasets without centralizing or sharing patient data.   This work presents a practical FL framework that supports geographically distributed collaboration among AI healthcare researchers and facilitates the development of robust models for oral cancer screening. Client devices were interconnected via Tailscale to provide secure networking and real-time communication. We implemented the FL workflow using the Flower framework for server-side aggregation, while client deployment and orchestration were configured manually; no enterprise FL platforms were used. To support a smartphone-based screening application, we evaluated lightweight, mobile-friendly architectures including MobileNetV2, MobileNetV3Large, and MobileNetV4-Conv-Small (MNv4-Conv-S). Across the global lightweight models aggregated using FedAvg, the MNv4-Conv-S based global model (GM-V4) achieved the best performance, reaching an AUC of 0.929 and an accuracy of 87%", "5": "2026-08-22T19:13:11.931975"}
{"0": 111, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19430v1", "3": "HARP: Hierarchical Adaptive Ranking with Preference-Adaptive Fusion for Query-Based CVE Prioritization", "4": "HARP: Hierarchical Adaptive Ranking with Preference-Adaptive Fusion for Query-Based CVE Prioritization. Authors: Haochen Liu, Zhengzhang Chen, Haoyu Wang, Yanchi Liu, Jundong Li. Categories: cs.IR cs.CL cs.CR. Vulnerability prioritization is inherently preference dependent, since the same CVE can receive different remediation priority under different operational preference scenarios. Existing scoring systems and ranking methods typically assume a fixed criterion. In practice, organizations already operate under a preference scenario, but this preference is often implicit and difficult to express as a written prompt instruction, while triage queries usually do not encode it. Past validated triage cases under the current scenario are more readily available. We study query-based CVE prioritization in this setting and propose HARP, a graph-grounded multi-view framework that ranks candidates from a natural-language query together with a support bank of historical labeled examples from the current preference scenario, without requiring an explicit textual summary of that scenario. HARP retrieves evidence from a vulnerability knowledge graph, scores candidates with policy-conditioned global, enterprise, and user views, and fits view-fusion weights from sampled supports. Experiments across three preference scenarios and multiple backbone LLMs show that HARP outperforms multiple baselines, expressing our method's effectiveness.", "5": "2026-08-22T19:13:11.936551"}
{"0": 112, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19401v1", "3": "Redactable blockchains and polynomial equations", "4": "Redactable blockchains and polynomial equations. Authors: Alexander Demin, Alexey Ovchinnikov, Vladimir Shpilrain. Categories: cs.CR cs.NI. We develop new tools for constructing redactable authenticated data structures with post-quantum security. In our construction, inverting the proposed one-way function means solving a polynomial equation (or a system of polynomial equations) in more than one variable. This is presently considered quantum-safe, i.e., there is no known quantum algorithm that could solve this problem efficiently if parameters are chosen wisely.", "5": "2026-08-22T19:13:11.942138"}
{"0": 113, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19387v1", "3": "Aray: Deterministic-First Synthesis of Benign Artifacts for YARA Validation", "4": "Aray: Deterministic-First Synthesis of Benign Artifacts for YARA Validation. Authors: Emanuel C. A. Valente, Louren\u00e7o A. P. J\u00fanior, Leonardo Gon\u00e7alves Chahud, J\u00falio Cezar Estrella, Marcus Botacin. Categories: cs.CR. A YARA rule is easy to distribute, but the malware sample used to demonstrate a positive match is not. This complicates storage, continuous integration, disaster-recovery exercises, and reproducible scanner validation. Constructing a replacement fixture requires more than embedding literals: YARA conditions can combine alternatives, counts, offsets, integer reads, and executable-container constraints, while the resulting file should not reproduce malware behavior. Positive validation is existential: it requires one file-level member of a rule's match set, not reconstruction of the originating sample. We present Aray, a deterministic-first YARA interpreter and positive-fixture synthesizer. Models may propose constructive normalizations or typed extraction fallbacks, but never backend source or binary structure. Conventional code validates normalized rules, derives string and integer witnesses, and performs extraction, routing, collision-checked layout, and ELF, PE, or generic serialization. Only residual normalization semantics reach a bounded model judge. We evaluated Aray over 416 public-rule entries. Normalization accepted 182 entries without model assistance and 234 after model normalization. Constructibility preflight admitted 406 entries, and every admitted fixture matched its upstream original rule. This yields 406/416 (97.6%) overall and 406/406 among constructible rules, with ten expected preflight dispositions and no scanner mismatches or construction failures. An unreachable endpoint confirmed zero model invocations during realization. The original-rule oracle validates generated fixtures against their source rules; proving implication for all possible files is a separate, stronger objective. Two anchored-regex failures were repaired before the final run, so these are post-fix systems results, not a held-out estimate.", "5": "2026-08-22T19:13:11.947321"}
{"0": 114, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19369v1", "3": "Linguistic Holonomy and Statistical Watermarks: Inner Geometry of Meaning-Preserving Transformations", "4": "Linguistic Holonomy and Statistical Watermarks: Inner Geometry of Meaning-Preserving Transformations. Authors: Daniele Corradetti. Categories: cs.CL cs.CR math.DG. Statistical watermarks for language models live in the freedom of the signifier: they choose among tokens that are nearly equivalent in meaning, and they are therefore eroded by exactly those transformations which move the form of a text while leaving its content in place. The literature measures such transformations by their endpoint, through the semantic similarity between the original and the rewritten text. We show that the endpoint is the wrong statistic. Adapting the formalism of linguistic loops, we prove that the invariant of a chain of meaning-preserving transformations factorises canonically into an endpoint part and a holonomy in the stabiliser of the initial state, the second of which the semantic deficit cannot see; the loop rotation is parallel transport on the unit sphere of the embedding space, so that the analogy with the Wilson loop becomes a theorem rather than a figure of speech. On the side of the detector we prove an exact identity: the residual statistic is proportional to the number of positions whose seeding window survived intact, from which the decay law $\u03c1^{h+1}$ follows as the independent-edit corollary. The identity has a disconcerting consequence, which we confirm to three decimal places: at one and the same retention rate the surviving signal may be one half of the original, one quarter of it, or exactly nothing, according only to where the edits fall.", "5": "2026-08-22T19:13:11.953255"}
{"0": 115, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19191v1", "3": "The Structured Totient Preimage Problem: Reconstruction, Collisions, and Cryptographic Implications", "4": "The Structured Totient Preimage Problem: Reconstruction, Collisions, and Cryptographic Implications. Authors: Luis Adri\u00e1n Lizama-P\u00e9rez. Categories: cs.CR. We define and study the Structured Totient Preimage (STP) problem as a restricted reconstruction relation with a direct cryptographic motivation. Let $p_1,\\ldots,p_k$ be distinct primes of the same bit length and reveal only $x=\\prod_{i=1}^k(p_i-1)$. Given $(x,\u03bb,k)$, STP asks for any set of $k$ distinct $\u03bb$-bit primes satisfying this product. The relation is efficiently verifiable, but its reconstruction complexity is not known. We establish three concrete results. First, for factored $x$ we derive the exact number of ordered exponent allocations and a bound showing that direct reconstruction is polynomial for fixed $k$ when $\u03a9(x)=O(\\log\u03bb)$; this rules out that regime as a basis for a strong hardness claim. Second, we give exhaustive algorithms for reconstruction and collision analysis. Third, we exhaustively evaluate 28 parameter pairs, with $2\\leq k\\leq5$, up to $\u03bb=16$ for pairs and 4,588,935 prime sets in the largest census. The data quantify non-injectivity through collision participation, maximum multiplicity, and conditional ambiguity in bits. These results isolate STP from general inverse-totient computation and motivate a Structured Totient Preimage Assumption for explicitly growing parameter families. Under such an assumption, STP becomes a candidate preimage-resistant relation whose implications for commitments, proofs of knowledge of multiplicative witnesses, and authentication can be stated precisely. The paper establishes the computational foundation and parameter constraints for those constructions; it does not claim a security reduction or post-quantum hardness.", "5": "2026-08-22T19:13:11.959230"}
{"0": 116, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19190v1", "3": "SiNMULI: Novel Signed Network Approach for Malicious URL Identification", "4": "SiNMULI: Novel Signed Network Approach for Malicious URL Identification. Authors: Avijit Gayen, Sayan Mondal, Angshuman Jana. Categories: cs.CR cs.NI cs.SE cs.SI eess.SY. In today's era of rapid advancements in artificial intelligence, computer security and online safeguarding measures have undergone significant improvements. However, malicious websites continue to facilitate the spread of phishing schemes, fraudulent activities and unsolicited communications. Conventional methodologies in machine learning, deep learning and counterfeit website detection predominantly depend on static data analysis, which frequently proves ineffective against the evolving nature of malicious online entities. In response to these challenges, in this work, we propose a signed network-based approach for malicious URL identification, SiNMULI. We introduce an innovative framework that conceptualises the identification of harmful URLs as a signed network-based binary classification problem strongly rooted in the fundamental principles of social network analysis and social balance theory. In this approach, a signed network is constructed based on the backlinks, i.e., external hyperlinks of URLs, wherein each node symbolises a URL and the hyperlinks function as signed edges. Utilising a balance-theoretic inference mechanism, our methodology propagates edge signs and classifies unlabeled domains by employing a 51% majority rule across incoming links. Experimental results on this real-world dataset demonstrate that SiNMULI achieves 99.89% accuracy, 99.62% precision, and 99.80% F1-score, outperforming traditional ML and deep learning baseline models. Beyond high accuracy, SiNMULI offers interpretability, resilience against adversarial obfuscation, and independence from training data, making it a lightweight and scalable solution for real-world cyber defence.", "5": "2026-08-22T19:13:11.966625"}
{"0": 117, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19161v1", "3": "Beyond the Transcript: Detecting Covert Co ordination in Latent Multi-Agent Communication", "4": "Beyond the Transcript: Detecting Covert Co ordination in Latent Multi-Agent Communication. Authors: Ramneet Kaur, Pradyumna Chari, Ramesh Raskar, Jugad Singh, Sumit Kumar Jha. Categories: cs.AI cs.CR. Language-model agents can communicate through continuous hidden states that are invisible in public transcripts, creating opportunities for covert harmful coordination. We introduce Verifiable Latent Alignments (VLA), an activation-aware framework for monitoring and steering these private communication channels. For every monitored decision, VLA links the private latent-state record and channel status to the resulting public action using a shared event identifier, enabling matched causal analysis. Our first contribution is a neutral-only three-layer monitor combining representation anomaly detection, counterfactual action-distribution influence, and sparse-autoencoder interpretation support. Our second contribution is a steerability framework spanning black-box behavioral instructions and white-box matched-neutral counterfactuals. Our third contribution is an evaluation on a controlled multi-agent auction benchmark covering homogeneous and heterogeneous model pairs, many-agent scalability, and intervention effectiveness. The sequential monitor achieves mean area under the receiver operating characteristic curve (AUROC) of 0.993 for homogeneous agents and 0.854 for heterogeneous pairs when text- and latent-collusion rows are pooled as positives. In Qwen3-0.6B auctions with 25-100 bidders, monitoring requires only a small normalized load relative to all possible directed pairs, while full white-box steering achieves 100% bid-distribution recovery and reduces collusive low-bid behavior by 47.3 percentage points. Because full white-box steering replays the matched neutral counterfactual, its exact recovery is a sanity check by construction. Overall, the controlled study shows that the evaluated private channel attacks can be monitored without training the primary monitor on attack examples and mitigated when matched counterfactual access is available.", "5": "2026-08-22T19:13:11.970483"}
{"0": 118, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19155v1", "3": "FedGuard-DC: Privacy-Preserving Federated Load Forecasting and Cyber-Attack Detection for Data-Center Loads in Transmission Systems", "4": "FedGuard-DC: Privacy-Preserving Federated Load Forecasting and Cyber-Attack Detection for Data-Center Loads in Transmission Systems. Authors: Md Kibria Saroare, Md Rubel Ahmed. Categories: cs.CR. The rapid growth of large data-center (DC) loads is creating new challenges for power-system visibility, privacy, and cyber-physical security. System operators need accurate short-term information about these fast-varying loads, while DC operators may avoid sharing raw megawatt measurements because they can reveal sensitive workload and utilization patterns. This paper presents FedGuard-DC, a federated learning (FL) framework for privacy-preserving DC load forecasting and local false-data-injection attack (FDIA) detection. Each DC trains a dual-head model on its own measurements, where a shared encoder supports both a forecasting head and a reconstruction head. A calibrated anomaly score combines forecast residual and reconstruction error to detect corrupted measurements locally. Raw measurements and absolute MW demand remain at each DC, while only model updates are shared with the global controller. Optional differential privacy and robust trimmed-mean aggregation are included to evaluate privacy-utility behavior and poisoned-client resilience. The framework is validated using EMT simulation data from four large DC loads rated between 150 and 350 MW integrated into the IEEE 39-bus New England system. Results show a 0.5 s-ahead normalized forecast RMSE of 0.023-0.038 pu, compared with 0.32-0.34 pu for persistence. FedGuard-DC detects FDIA with ROC-AUC of 0.979, F1 = 0.930, and precision of 0.988, while robust aggregation reduces the poisoned-client RMSE impact from 0.042 to 0.035 pu.", "5": "2026-08-22T19:13:11.976603"}
{"0": 119, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19135v1", "3": "Autonomous Cyber Defense in Connected Vehicles: A Multi-Agent Approach to V2X Security", "4": "Autonomous Cyber Defense in Connected Vehicles: A Multi-Agent Approach to V2X Security. Authors: Krishna Teja Medam. Categories: cs.CR cs.DC cs.MA cs.NI. A connected vehicle has roughly 100 milliseconds to decide whether an incoming Basic Safety Message is real or fabricated. If a false emergency braking alert reaches the planning pipeline in time, the car brakes - a safety failure triggered by a security failure. Existing intrusion detection systems are not designed to handle that coupling. They operate per vehicle, per message, with static rules - blind to attack patterns that only emerge across a fleet or over time, and blind to the fundamental tension between dropping a suspicious message and dropping a real emergency alert. We propose a three-tier multi-agent architecture that treats this timing constraint as a hard design requirement, not a performance target. At the vehicle level, an onboard agent classifies each incoming V2X message into one of four actions - Accept, Drop, Quarantine, or Escalate - within a 10-millisecond budget, deliberately biased toward Escalate when uncertain, passing ambiguous cases to the roadside edge agent rather than risking a dropped legitimate alert. The edge agent operates across a roadside unit zone with a 50-millisecond budget, fusing threat assessments from multiple vehicles and resolving safety-security conflicts using complementary sensor observations. The cloud tier refines detection models through Byzantine fault-tolerant federated learning and redistributes updated weights to the fleet. Every timing constraint derives directly from the 100-millisecond Basic Safety Message cycles mandated by SAE J2735 and ETSI EN 302 637-2. No existing framework simultaneously assigns standards-grounded latency budgets to all three deployment tiers while treating safety-security conflict resolution as a first-class design constraint. Remaining open problems - adversarial poisoning at the edge and the absence of regulatory frameworks for autonomous security response - are discussed as future work.", "5": "2026-08-22T19:13:11.982407"}
{"0": 120, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19302v1", "3": "ABEAT: Efficient and Anonymous Encryption for ABE-based Dynamic Group Communication", "4": "ABEAT: Efficient and Anonymous Encryption for ABE-based Dynamic Group Communication. Authors: Hongmiao Yu, Silas Richelson, Jiachen Chen, K. K. Ramakrishnan. Categories: cs.CR cs.NI. Confidential communication among a dynamic group of participants that ensures flexible and efficient many-to- many communication is highly desired capability. We leverage attribute-based encryption (ABE) for confidential group communication and enhance it by a graph-based namespace to create an efficient framework that allows groups to be formed and changed dynamically. In this paper, we focus on the important additional need to maintain the anonymity of recipients of a message, when using ABE for group communication for a variety of usage scenarios (e.g., emergency response).   We propose ABEAT, an efficient and anonymous dynamic group communication system that also minimizes overhead on receivers who are not the intended recipients of a message. In ABEAT, we propose a new anonymous KP-ABE approach to maintain recipient anonymity. ABEAT hides the clear attribute in the ciphertext of KP-ABE, and also prevents several attacks that seek to break anonymity. ABEAT provides fast recipient verification, dramatically lowering the decryption overhead for non-recipients by more than a factor of 90 versus the current state of the art such as hidden vector encryption (HVE). In fact, it is even 40% less than FABEO, which offers no anonymity.", "5": "2026-08-22T19:13:11.988616"}
{"0": 121, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20340v1", "3": "Group-theoretic treatment of strong light-matter coupling with an arbitrary number of excitations", "4": "Group-theoretic treatment of strong light-matter coupling with an arbitrary number of excitations. Authors: Antti Peltola, Olli Siltanen, Kimmo Luoma, Konstantinos S. Daskalakis. Categories: quant-ph. Strong light-matter interactions in optical microcavities give rise to hybrid light-matter states known as polaritons. While actively used in modern technologies, theoretical descriptions of such systems are often restricted to the single-excitation case, limiting their ability to capture many-excitation physics and hindering further technological advancements. Here, by exploiting the combinatorial structure of quantum emitters, we investigate the Tavis-Cummings model with arbitrary number of excitations. We derive the structure and properties of its eigensystem and identify allowed radiative transitions in systems of realistic size scales. Our work reveals new behavior inaccessible to the few-excitation regime, while also providing a framework to reduce the computational complexity of similar systems with exponentially growing Hilbert spaces.", "5": "2026-08-22T19:13:16.184885"}
{"0": 124, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20335v1", "3": "4DAnyone: Create Anyone in 4D from a Casual Monocular Video", "4": "4DAnyone: Create Anyone in 4D from a Casual Monocular Video. Authors: Yudong Jin, Tao Xie, Qihang Zhang, Zehong Shen, Zhen Xu. Categories: cs.CV. We present 4DAnyone, a framework for reconstructing 4D humans from an uncalibrated monocular video by generating reconstruction-grade multiview-consistent videos and lifting them into 4D Gaussian Splatting (4DGS). Existing camera-controlled video diffusion models synthesize plausible novel-view videos but fail to maintain consistency when scaled to the tens of target views required for 4DGS reconstruction. We identify this failure as a bounded-attention-context problem: when target views exceed the capacity of a single DiT forward pass, they must be split into groups, exposing two coupled bottlenecks. On the reference-context side, conditioning on all previously generated views grows as $O(N)$, weakening cross-view appearance guidance. On the target-context side, disjoint groups cannot directly exchange information, causing global structural drift. 4DAnyone addresses both bottlenecks with two complementary designs: Reference Context Packing (RCP) compresses growing reference views into a fixed-length mixed-resolution context with $O(1)$ reference-context complexity, while Target Context Routing (TCR) rotates target-view groupings during denoising to share context across groups at high-noise steps and stabilize details at low-noise steps. We further build the MVGameHuman dataset using our in-house game engine and combine it with light-stage and in-the-wild video datasets for training. Experiments on DNA-Rendering and DyMVHumans show that 4DAnyone outperforms prior methods in both novel-view video quality and downstream 4DGS reconstruction, with robust in-the-wild generalization. See our project page for video results and source code: https://4danyone.github.io.", "5": "2026-08-22T19:13:16.201942"}
{"0": 125, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20336v1", "3": "WithEveryone: Unified Planning and Identity Grounding for Group Image Generation", "4": "WithEveryone: Unified Planning and Identity Grounding for Group Image Generation. Authors: Hengyuan Xu, Qixun Wang, Yiji Cheng, Miles Yang, Zhao Zhong. Categories: cs.CV. Identity-preserving image generation becomes increasingly unreliable when a scene must contain many specified people. Beyond retaining each identity, the model must bind every reference to a distinct person and location, while training-time identity losses must establish correspondence among several noisy predicted faces. We introduce WithEveryone, a unified framework for generating group images up to ten reference identities. WithEveryone injects each selected identity as an addressed token, predicts a structured identity--layout plan, and renders the plan as a visual condition. Its key objective, Layout-Grounded ID Loss, uses annotated face regions to supervise the intended identities directly, avoiding unstable embedding-based face matching; ID Representation Forcing additionally trains a prediction for each identity before image synthesis. On an identity-disjoint benchmark, WithEveryone achieves the highest target-context identity similarity, improving face similarity from 0.462 for GPT-Image-2 to 0.499, while reducing copy-paste artifacts from 0.169 to 0.055. It further covers 97.3\\% of the requested identities with a duplicate rate of only 2.8\\%. These results show that explicit identity--layout grounding enables identity-preserving generation to scale to larger groups without relying on direct reference-face copying.", "5": "2026-08-22T19:13:16.208734"}
{"0": 126, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20334v1", "3": "Swift-Image: Exploring the Performance Frontier of Compact Unified Image Generation Models", "4": "Swift-Image: Exploring the Performance Frontier of Compact Unified Image Generation Models. Authors: Taihang Hu, Zhao Wang, Zuan Gao, Tao Liu, Hao Yan. Categories: cs.CV. We present Swift-Image, a compact unified model for text-to-image generation, single-image editing, and multi-image editing. Our goal is to explore how far a relatively small visual generator can be pushed through systematic training engineering under a constrained computational budget. Swift-Image adopts an efficient 6B single-stream DiT and a progressive training pipeline that evolves from broad semantic coverage to higher resolution, stronger visual quality, and unified generation-editing supervision. For post-training, we employ parallel expert reinforcement learning followed by multi-teacher on-policy distillation to alleviate interference among heterogeneous objectives. We further decouple high-level reasoning from pixel-level rendering with a Prompt Enhancer that translates user requests into generator-aligned visual specifications. For efficient deployment, structural pruning and few-step distillation produce 3B and accelerated variants. Swift-Image achieves leading aggregate performance among evaluated open-source models with only 6B parameters and 243K GPU training hours; the compressed 3B model incurs nearly no loss, while few-step distillation further improves aggregate editing performance with substantially fewer sampling steps. Our study also summarizes practical lessons for architecture, data curriculum, post-training, prompt enhancement, and model compression.", "5": "2026-08-22T19:13:16.216571"}
{"0": 128, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20329v1", "3": "Production of quarkonium pairs via the fragmentation mechanism", "4": "Production of quarkonium pairs via the fragmentation mechanism. Authors: Franco Barattini, Benjamin Guiot, Marat Siddikov. Categories: hep-ph. In this manuscript, we analyze the contribution of fragmentation mechanisms to the inclusive hadroproduction of heavy quarkonium pairs, taking into account both single- and double-hadron (dihadron) fragmentation channels. To estimate the contribution of the latter mechanism, we construct a microscopic perturbative model of the dihadron fragmentation function, which is valid in the limit of large invariant masses of the heavy quarkonium pair. Using the Color Glass Condensate framework to evaluate the $Q\\bar{Q}$ production amplitude, we find that in LHC kinematics, the single fragmentation of heavy quarks into quarkonium is only a minor correction, whereas dihadron fragmentation can provide a sizable contribution, on par with the contributions of single- and double-parton scattering.", "5": "2026-08-22T19:13:16.226480"}
{"0": 130, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20323v1", "3": "Spin-torque microwave detectors of positive rectangular pulse signals", "4": "Spin-torque microwave detectors of positive rectangular pulse signals. Authors: V. Prokopenko, O. Shtanko, I. Sotnyk, O. Prokopenko. Categories: physics.app-ph cond-mat.mes-hall. We analyze the performance of a spin-torque microwave detector (STMD) driven by positive rectangular current pulses $I(t)$ of various amplitudes $I_0$, durations $\u03c4$, and repetition periods $T$ and reveal two distinct regimes of STMD operation. In the first (linear) regime, the time-averaged voltage across the detector, $U_{\\rm dc}$, changes linearly with the pulse amplitude $I_0$ and depends on the ratio $\u03c4/T$: $U_{\\rm dc} \\sim I_0 (\u03c4/T)$. This regime is observed for a wide range of pulse amplitudes $I_0$ in the case of in-plane (IP) magnetization dynamics and for rather small pulse amplitudes $I_0 \\le I_{\\rm th}$ in an STMD with out-of-plane (OOP) magnetization dynamics. The other (nonlinear) regime is characterized by voltage jumps and drops and is observed only in a structure with OOP magnetization dynamics for input pulses with short repetition periods and large amplitudes $I_0 \\ge I_{\\rm th}$. We believe that the linear regime of STMD operation can be used to unambiguously detect input pulse parameters, which could be important for the development and optimization of spintronic devices capable of detecting and processing non-harmonic (e.g., digital) microwave signals.", "5": "2026-08-22T19:13:16.238046"}
{"0": 131, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20321v1", "3": "Large Sample Properties of Higher Order Markov Models", "4": "Large Sample Properties of Higher Order Markov Models. Authors: Tuhin Majumder, Donald E. K. Martin, Soumendra N. Lahiri. Categories: math.ST math.PR. We study large-sample properties of higher-order Markov chains on a finite alphabet $\u03a3$ when the order $m_n$ is allowed to grow with the sequence length $n$. By embedding the process into a first-order chain on $\u03a3^{m_n}$ and exploiting return-time decompositions, we establish a central limit theorem for additive functionals $\\sum_{t}\\! g_n(Y_t^{(n)})$ under natural ergodicity and sparsity conditions. The normalization involves the stationary return time to a suitably chosen state and accommodates triangular arrays with $m_n\\!\\to\\!\\infty$ and $m_n/n\\!\\to\\!0$. We further illustrate the assumptions in a binary variable length Markov chain (VLMC), deriving explicit lower bounds on stationary masses that yield a concrete growth regime (e.g., $m_n\\log m_n/n \\to 0$) ensuring the CLT. These results provide asymptotic foundations for inference in sparse/partitioned higher-order models; including VLMCs and sparse Markov models (SMMs) where the effective dimensionality grows with the sample size.", "5": "2026-08-22T19:13:16.242888"}
{"0": 135, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20317v1", "3": "Projecting BrowseComp-Plus onto ClimbMix: Toward More Realistic Corpora for Agentic Search", "4": "Projecting BrowseComp-Plus onto ClimbMix: Toward More Realistic Corpora for Agentic Search. Authors: Sahel Sharifymoghaddam, Lingwei Gu, Yijun Ge, Jimmy Lin. Categories: cs.IR. The BrowseComp-Plus benchmark disentangled the evaluation of agentic search by replacing opaque web search with a fixed corpus, so that an agent's role can be separated from the retriever's. That corpus, however, holds only about 100K documents and was assembled from the supporting documents of the benchmark's own queries plus mined hard negatives, so the evidence and the distractors were both selected per query. We introduce $\\text{BrowseComp-Plus}_{\\text{CM}}$, which keeps the BrowseComp-Plus questions but relocates their evidence to ClimbMix, a 400B-token, 553M-document mixture of web text released by NVIDIA for pre-training language models and built without reference to any benchmark. Our main contribution is the projection pipeline that makes this possible: it decomposes each question into atomic reasoning hops and grounds every hop in the new corpus, retaining a question only when automatic verification, an independent agent, and human review all confirm that every hop is supported. The pipeline is dataset-agnostic and applies to any benchmark whose questions decompose into verifiable facts. Applied to the 830 BrowseComp-Plus test questions, our pipeline yields 57 fully grounded questions with question-level relevance judgments. Projection shifts the difficulty onto retrieval, as the strongest agent we evaluate loses five points of answer accuracy but sees its evidence recall fall from 84.3% to 21.4% while issuing 63% more search calls. As the first of a series of projections, we release the pipeline, the benchmark, and our analyses at https://github.com/castorini/cmass.", "5": "2026-08-22T19:13:16.262676"}
{"0": 139, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20312v1", "3": "Inter-X++: A Comprehensive Benchmark for Multimodal Human-Human Interaction Analysis", "4": "Inter-X++: A Comprehensive Benchmark for Multimodal Human-Human Interaction Analysis. Authors: Liang Xu, Chengqun Yang, Zili Lin, Xintao Lv, Yichao Yan. Categories: cs.CV. The capability to perceive and synthesize human-human interactions is fundamental to developing intelligent digital human systems. However, existing datasets and modeling approaches are fundamentally constrained by low-fidelity kinematics, the omission of dexterous hand gestures and a severe lack of rich multimodal annotations. Furthermore, fragmented interaction representations and inconsistent evaluation protocols also impede fair and rigorous benchmarking. To systematically address these bottlenecks, we present Inter-X++, a comprehensive and large-scale benchmark designed to empower versatile HHI analysis. Captured via a novel hybrid motion capture system, Inter-X++ provides 11,388 high-fidelity interaction sequences and over 8.1M frames, featuring precise whole-body movements and detailed finger articulations. Meanwhile, we enrich the data foundation with multifaceted annotations, including hierarchical fine-grained textual descriptions, interaction categories, causal interaction orders, the relationship and personality of the subjects, as well as vertex-level contact maps and physically regularized constraints. Leveraging these elaborate annotations, we formulate a unified testing ground comprising four categories of downstream tasks that symmetrically span both generative and perceptive paradigms. To eliminate benchmarking ambiguities, we systematically standardize the interaction representations and evaluation protocols. Finally, we go beyond dataset construction to propose OpenHHI, a single and unified HHI representation and modeling framework that jointly optimizes interaction reconstruction and semantic understanding. Extensive experiments reveal that OpenHHI achieves state-of-the-art performance on both generation and perception tasks. This definitively proves that our unified representation successfully bridges interaction understanding and generation simultaneously.", "5": "2026-08-22T19:13:16.283896"}
{"0": 140, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20310v1", "3": "Signatures of a light-induced exciton condensate exhibiting BEC-BCS crossover", "4": "Signatures of a light-induced exciton condensate exhibiting BEC-BCS crossover. Authors: Khanh Duy Nguyen, Gabriele Berruto, Yunhe Bai, Thomas Marchese, Woojoo Lee. Categories: cond-mat.mes-hall cond-mat.quant-gas cond-mat.str-el. Exciton condensates provide a platform to study quasiparticle pairing, Bose-Einstein condensation-Bardeen-Cooper-Schrieffer (BEC-BCS) crossover, and excitonic topological phenomena. Achieving a nonequilibrium exciton condensate allows the ultimate tunability of these emergent phenomena. Yet, evidence of a light-induced, nonequilibrium exciton condensate and its BEC-BCS crossover remains elusive. Here, we use time- and angle-resolved photoemission spectroscopy to demonstrate signatures of a non-equilibrium exciton condensate and its BEC-BCS crossover in monolayer MnBi2Te4. Following optical excitation, a distinctive hole-like dispersion representing excitons emerges and persists for >20 ps. Strikingly, energy-domain sharpening in the valence band occurs 2 ps after time zero and exhibits a sharp onset at a threshold pump fluence of 0.84 mJ/cm2. The delayed and strongly nonlinear response is difficult to reconcile with transient field effects or conventional carrier-induced band shifts but is consistent with a model of exciton condensation governed by a Berezinskii-Kosterlitz-Thouless transition. The estimated threshold exciton density agrees quantitatively with the Nelson-Kosterlitz critical density. At higher fluences, the exciton feature develops a camel-back-shaped dispersion, consistent with the BEC-BCS crossover in the condensate framework. Our work establishes ultrathin MnBi2Te4 as a model system for studying nonequilibrium exciton condensates with a connection to superconductivity and exciton-driven topological phases.", "5": "2026-08-22T19:13:16.290319"}
{"0": 141, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20308v1", "3": "DreamHand: Repurposing Video Diffusion Models for Occlusion-Robust Egocentric 3D Hand Motion Recovery", "4": "DreamHand: Repurposing Video Diffusion Models for Occlusion-Robust Egocentric 3D Hand Motion Recovery. Authors: Yufei Liu, Xixi Wang, Hao Li, Ganlong Zhao, Kaitong Cai. Categories: cs.CV. Egocentric video offers scalable manipulation data for embodied AI, yet recovering metric 3D hand trajectories remains challenging due to severe object occlusion and frequent out-of-sight gaps. Existing single-frame and windowed temporal regressors fail when hand shortly leaves the frame, while recent video diffusion models (VDMs) rely on heavy, stochastic multi-step sampling as pixel-space renderers. We instead repurpose VDM into a deterministic geometry encoder. A single forward pass over the clean latent exposes scene content beyond current observations, including occluded and out-of-sight hands. We introduce DreamHand, an offline clip-level framework that extracts features via a Deterministic Clean-Latent Encoder and decodes them with a Bidirectional Spatiotemporal Decoder. DreamHand recovers continuous bimanual trajectories with metric placement and no external detector, while a Ray-Based Camera Solver supports a second configuration that needs no test-time camera intrinsics. Across five egocentric benchmarks, DreamHand sets a new state of the art, cutting MPJPE-p by 30% on occlusion-heavy ARCTIC and 40% on HOT3D. These gains reach 46%-61% once out-of-sight hands are included in the evaluation, offering a scalable path from everyday human video to robot manipulation data.", "5": "2026-08-22T19:13:16.296609"}
{"0": 142, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20306v1", "3": "Computational Methods of Wave Propagation for Semiclassical Models of High Harmonic Generation in Bulk Solids", "4": "Computational Methods of Wave Propagation for Semiclassical Models of High Harmonic Generation in Bulk Solids. Authors: Ava N. Hejazi, Nicholas Karpowicz, Gregory D. Scholes, Julia M. Mikhailova. Categories: physics.optics physics.chem-ph. We present a theoretical framework for self consistent treatment of nonlinear light-matter interactions in the ultra-fast strong-field regime based on numerical solution of Maxwell's equations and semiconductor Bloch equations. This framework is shown to describe high-order harmonic generation and propagation in bulk semiconductors, investigating differences in reflected and transmitted harmonic spectra due to propagation effects. We show that the propagation of the combined field of the driving laser pulse and generated harmonics in a bulk semiconductor significantly modifies the harmonic spectra, affecting interpretation of experimental results relating the transmitted harmonic spectra to the underlying electronic structure of the material. This model allows the self-consistent description of strong-field light-matter interactions in the non-perturbative regime, opening the way to explore the transition between fully classical and quantum regimes of interaction, tunneling and multiphoton regimes of material ionization, and perturbative and non-perturbative regimes of harmonic generation in bulk materials.", "5": "2026-08-22T19:13:16.301192"}
{"0": 143, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20298v1", "3": "Zero-Sum Power Factor Games", "4": "Zero-Sum Power Factor Games. Authors: Cameron Khanpour, Samuel Talkington, Mathieu Dahan, Daniel K. Molzahn. Categories: eess.SY math.OC. Variable active power injections arising from device behavior or compromised dispatch complicate voltage regulation in electric power networks with distributed energy resources (DERs). An operator can limit the resulting voltage deviations by remotely selecting DER reactive power parameters before observing the active power injections. IEEE Standard 1547-2018 specifies constant power factor as one such control mode, coupling each device's reactive power to its realized active power. Using a linear voltage model, we formulate the operator's decision as a robust minimax problem in which the operator minimizes the largest feasible aggregate voltage deviation. We solve this problem by expressing the power factor decisions through continuous reactive to active power ratios and exactly decomposing the payoff according to the signs of the voltage deviations. When every feasible voltage residual remains on its initial side of nominal, the resulting ratios cancel each injection's contribution and yield a closed form minimax strategy. We identify realistic DER ratings for which this strategy applies and quantify the regulation capacity lost under restricted power factor ranges. Numerical tests check the cancellation computation, solve the complete minimax problem directly at a representative DER rating, and compare the linear voltage predictions with nonlinear AC power flow.", "5": "2026-08-22T19:13:16.309095"}
{"0": 144, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20296v1", "3": "A blind spot in transverse BAO calibration", "4": "A blind spot in transverse BAO calibration. Authors: Domenico Sapone. Categories: astro-ph.CO. Transverse baryon acoustic oscillation (BAO) measurements are increasingly used for cosmological inference, and carry a calibration that no such inference can constrain. A constant error in the transverse BAO scale is exactly degenerate with the combination $r_{\\rm d} h$: it leaves the goodness of fit unchanged and the recovered parameters plausible, and is therefore invisible to any analysis that uses these measurements alone. The radial BAO sector removes this degeneracy, supplying $D_{\\rm M}/r_{\\rm d}$ by integration without reference to $H_0$ or to any model for the expansion rate. In flat Friedmann--Lema\u00eetre--Robertson--Walker (FLRW) geometry the relation between the two sectors is an identity, so the sound horizon and the dark-energy equation of state cancel as well. An integrated form of this identity reduces the consistency test to a straight line, whose slope measures a relative transverse calibration $\\varepsilon$. A departure from $\\varepsilon = 1$ cannot be produced by any dark-energy model, nor by spatial curvature: it indicates an inconsistency in the measurement chain rather than in the cosmology. We apply the test to the two main SDSS transverse BAO compilations, which give $\\varepsilon = 1.073 \\pm 0.021$ and $1.021 \\pm 0.029$. The first differs from unity at $3.8\u03c3$ using the published independent errors, while the second is consistent with unity. The two compilations themselves differ by $(5.4 \\pm 1.4)\\%$, or $3.9\u03c3$, and the offset is constant in redshift. The test provides a direct diagnostic for current and future angular BAO measurements.", "5": "2026-08-22T19:13:16.312817"}
{"0": 146, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20288v1", "3": "Transient Early Dark Energy-Like Dynamics as a Mechanism for Enhanced Early Structure Formation in the JWST Era", "4": "Transient Early Dark Energy-Like Dynamics as a Mechanism for Enhanced Early Structure Formation in the JWST Era. Authors: Abhik Bhattacharjee, Amlan Chakraborty, Subinoy Das, Anshuman Maharana, Priyank Parashari. Categories: astro-ph.CO. The discovery of massive galaxies at redshifts $z\\gtrsim10$ by the James Webb Space Telescope (JWST) has renewed interest in cosmological mechanisms capable of enhancing early structure formation while preserving the successful large-scale predictions of the standard $\u039b$CDM model. We investigate a phenomenological scenario in which an exotic dark matter species can undergo a transient early dark energy-like phase during the radiation-dominated era ($10^{-7}\\lesssim a\\lesssim10^{-5}$) before reverting to pressureless cold dark matter. We utilize the generalized dark matter framework to model this species, which is restricted to a sub-percent fraction of the total dark matter component by CMB, BAO and Type Ia supernova data. Its background and perturbation dynamics are characterized by a time-dependent equation of state, $w(a)$, and a time- as well as scale-dependent sound speed, $c_s^2(a,k)$. The temporary negative equation of state, combined with our phenomenological pressure-response prescription, induces a finite interval of negative effective sound speed squared. This triggers an instability-driven growth of density perturbations over a limited range of comoving scales, thereby enhancing the formation of early dark matter halos. We find that the enhanced halo abundance can substantially reduce the star-formation efficiencies required to reproduce the observed abundance of JWST galaxies relative to the standard $\u039b$CDM scenario, especially at higher redshifts. Our results demonstrate that transient early dark energy-like dynamics in a subdominant dark matter component provide a viable mechanism for enhancing early structure formation and offer a new framework for interpreting the abundance of high-redshift galaxies observed by JWST and future surveys.", "5": "2026-08-22T19:13:16.322621"}
{"0": 148, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20284v1", "3": "Towards Surgical World-Action Modeling: A Preliminary Joint Visual-Trajectory Forecasting for Surgical Motion Planning", "4": "Towards Surgical World-Action Modeling: A Preliminary Joint Visual-Trajectory Forecasting for Surgical Motion Planning. Authors: Weiliang Huang, Huanrong Liu, Bob Zhang, Qi Dou, Zhen Chen. Categories: cs.CV cs.RO. Reliable surgical planning requires models to anticipate not only how instruments will move, but also how the operative visual state will evolve together with such motion. Existing approaches typically treat future scene generation and instrument trajectory prediction as two separate tasks. Scene-only models cannot directly evaluate the accuracy of future instrument motion at the trajectory level, while trajectory-only models fail to capture the visual consequences of instrument movement, leaving the consistency between predicted trajectories and future scene evolution unaddressed. Jointly forecasting both provides a more complete account of surgical action-scene dynamics by enabling explicit trajectory-level evaluation while simultaneously modeling the corresponding visual evolution. To bridge this gap, we present a preliminary joint visual-trajectory world-action model that simultaneously forecasts future visual states and instrument trajectories from historical surgical observations. Specifically, we encode historical video frames and tool trajectories into latent representations, which are processed by a temporal-spatial encoder and subsequently decoded through separate visual-state and trajectory prediction heads. Based on this preliminary architecture, a chunked autoregressive rollout is repeatedly applied to predict fifteen future steps. The chunked strategy consistently outperforms direct one-shot prediction across all evaluated horizons, improving first-segment PSNR from 18.86 to 23.11 dB and reducing ADE from 45.77 to 22.22 pixels. These results demonstrate the initial feasibility of joint visual-motion forecasting. However, we observe progressive visual degradation and accumulated trajectory errors over longer prediction horizons, which remain important challenges for future surgical world-action modeling.", "5": "2026-08-22T19:13:16.331733"}
{"0": 149, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20283v1", "3": "Design and Empirical Evaluation of a Network-Centric, On-Premises Architecture for Earth Observation Data Access", "4": "Design and Empirical Evaluation of a Network-Centric, On-Premises Architecture for Earth Observation Data Access. Authors: Jo\u00e3o Pinelo, Jo\u00e3o Gon\u00e7alves, Denis Willett, Amit Ruhela, Derek Steinmoeller. Categories: cs.DC cs.NI cs.PF. Earth observation (EO) programmes generate data at volumes that exceed the transfer and storage capacity of most institutional networks. Public cloud platforms address this for well-resourced organisations, but institutions across the Atlantic basin face constraints in connectivity, sovereignty and funding that make on-premises infrastructure the only viable path. Cloud-native data formats enable efficient partial reads, yet their performance depends on the bandwidth of the underlying network fabric, a dependency rarely measured in isolation. This paper presents a replicable, network-centric architecture for on-premises EO data access, evaluated at its first operational deployment: the AIR Data Centre, founding node of the Atlantic Cloud. The system comprises a MinIO object storage cluster on a 100 GbE fabric, a PostGIS metadata catalogue and an OGC API-EDR access layer. We characterise the fabric under sustained parallel load, evaluate object storage throughput for EO-representative workloads, and compare measured performance against throttled baselines on identical hardware, isolating network bandwidth as the sole variable. Multi-site replication benchmarks with partner institutions characterise the federation primitive the model depends on. Network bandwidth is the dominant constraint on storage throughput for bulk EO data access up to a threshold; beyond it, endpoint memory topology rather than capacity governs how much bandwidth a system can use. For this hardware class that threshold lies above 10 Gbps per server. Below it, network capacity alone sets what the facility can deliver; above it, the return on further network investment depends on endpoint memory provisioning, which can be deferred and bought later.", "5": "2026-08-22T19:13:16.337303"}
{"0": 159, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20260v1", "3": "From Kriging to Spatial AI: Fifty Years of Spatial Statistics for Complex Dependent Data", "4": "From Kriging to Spatial AI: Fifty Years of Spatial Statistics for Complex Dependent Data. Authors: Montserrat Fuentes, Veronica B. Patterson. Categories: stat.ME. Spatial statistics has grown from kriging for spatial prediction into a broad framework for learning from complex dependent data. This article traces that development from random fields and spectral methods to Bayesian hierarchical models and scalable computation. It then connects these foundations to Spatial AI, where graph learning and neural networks are being adapted to spatially dependent data. The article introduces the main ideas behind kriging and nonstationarity and explains how data fusion and uncertainty quantification extend spatial inference to more complex settings. The central contribution is a unified account of how these developments lead naturally to new forms of Spatial AI. Rather than treating spatial statistics and machine learning as separate traditions, we show how both learn from dependence while preserving interpretable structure. We also examine how spatial geometry and physical knowledge can guide flexible representation learning and support scientifically meaningful prediction.", "5": "2026-08-22T19:13:20.385770"}
{"0": 160, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20253v1", "3": "GENIE: Generative Neural Inference for Epidemics", "4": "GENIE: Generative Neural Inference for Epidemics. Authors: Laura M. Guzm\u00e1n-Rinc\u00f3n, George R. E. Bradley, Joel Kandiah, Kyriakos Flouris, Pietro Li\u00f2. Categories: stat.ME q-bio.QM. The SARS-CoV-2 pandemic highlighted the ongoing risk infectious diseases pose to society and the value of reliable information on the likely future burden. When forecasting an epidemic at fine spatial resolution, traditionally used mechanistic compartmental model struggle to capture highly complex granular transmission dynamics, resulting in inaccurate and overconfident forecasts. However, detailed Agent-Based Models (ABMs), are challenging to calibrate and are too computationally expensive to use in real-time. Amortized simulation-based inference promises to overcome this difficulty by exploiting the power of machine learning (ML) to perform approximate forecasting at near-real-time using arbitrarily complex models of epidemics. In this work we introduce Generative Neural Inference for Epidemics (GENIE), a spatio-temporal ML-based framework for high-resolution forecasting of the burden of respiratory pathogens. GENIE is designed to reflect two key characteristics of outbreaks: (i) shared biological mechanisms across locations and (ii) location-specific characteristics affecting transmission dynamics. This results in the model architecture having two modules: (i) a Local Infection Encoder - which learns to represent disease dynamics shared across all locations and (ii) a Local Profile Encoder - which learns location-specific representations. Using simulations from a high-resolution spatio-temporal ABM, GENIE is trained to generate samples from an approximate posterior predictive distribution of future epidemic trajectories. Benchmarked against established statistical and ML models, GENIE demonstrates superior performance across a range of measures including the timing and magnitude of peak hospitalisations.", "5": "2026-08-22T19:13:20.391754"}
{"0": 161, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20251v1", "3": "Video2DoorTraversal: Push Door Traversal via Simulated Door Twins", "4": "Video2DoorTraversal: Push Door Traversal via Simulated Door Twins. Authors: Xincheng Tang, Yiji Chen, Youhan Xie, Wanyu Li, Zhengjie Shu. Categories: cs.RO. Door opening and traversal is a long-horizon loco-manipulation task that requires precise handle interaction and coordinated base-arm control. We present Video2DoorTraversal, a single-video real-to-sim-to-real framework for wheel-legged mobile manipulators. Given one RGB video of a real door, DoorTwin reconstructs an instance-aligned, articulated, and simulation-ready door twin with realistic geometry and appearance. A simulation-in-the-loop agent converts the recovered articulation into a parameterized skill program and iteratively refines failed rollouts to generate physically executable demonstrations. These demonstrations are used to train ArticuACT, a dual-depth policy that predicts coordinated base, arm, and gripper commands using robot-centric camera conditioning and interaction-aware supervision. With all perception and policy inference running onboard, the system achieves a 96.57% average success rate across five real doors and an 80.95% zero-shot success rate on structurally similar unseen doors, while completing the full approach, opening, and traversal sequence in approximately 13s on average. Project Page: https://video2doortraversal.github.io/.", "5": "2026-08-22T19:13:20.399883"}
{"0": 167, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20154v1", "3": "Artificial Intelligence for Workflow Analysis in Colorectal Surgery: A Multicentric, Cross-Procedural Development and Generalization Study", "4": "Artificial Intelligence for Workflow Analysis in Colorectal Surgery: A Multicentric, Cross-Procedural Development and Generalization Study. Authors: Pietro Mascagni, Julia Alekseenko, Pooja P Jain, Marta Goglia, Andrea Balla. Categories: cs.CV. Minimally invasive colorectal surgeries (MIS-CRS) are characterised by significant variability and inconsistent outcomes. ColoWorkflow, a tool for the video-based assessment (VBA) of MIS-CRS workflow, was recently validated. However, manual VBA is time-consuming, limiting implementation. This study presents AI-ColoWorkflow, a deep learning model for automated surgical workflow analysis across MIS-CRS. Operative videos of MIS-CRS were collected from 4 centres and a publicly available dataset. Phases and steps were manually annotated according to ColoWorkflow. A deep learning model combining a fine-tuned DINOv3 vision transformer for per-frame visual feature extraction with a hierarchical multi-stage temporal convolutional network was jointly optimized for phase and step recognition. The model trained on pooled multicentric data, namely AI-ColoWorkflow was compared against centre-specific and procedure-specific models on a held-out test set. The following metrics were used for evaluation: macro F1 score, balanced accuracy, precision, and recall. AI-ColoWorkflow achieved a macro F1 of 73.01% $\\pm$ 10.27 (balanced accuracy 73.43%) for phase recognition and 39.82% $\\pm$ 7.06 (balanced accuracy 38.65%) for step recognition. The global model outperformed centre- and procedure-specific models in most experiments except procedure-specific step recognition. In the generalization analysis, mean F1 was 48.42% for phase recognition. AI-ColoWorkflow can reliably recognize MIS-CRS phases. A single model trained on pooled, multicentric, multi-procedural data generalises at least as well as and often better than centre- or procedure-specific models for phase recognition in MIS-CRS, while procedure-specific step models retain advantages for certain procedure types, motivating hybrid training strategies for future surgical AI development.", "5": "2026-08-22T19:13:20.431602"}
{"0": 168, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20139v1", "3": "Reinforcement LearningtoHarness Approximation Errors for Long-Time QuantumSimulation", "4": "Reinforcement LearningtoHarness Approximation Errors for Long-Time QuantumSimulation. Authors: Yu-Bo Shi, Markus Heyl, Roderich Moessner, Marin Bukov, Hongzheng Zhao. Categories: quant-ph cond-mat.dis-nn cond-mat.stat-mech cond-mat.str-el. Accurate digital quantum simulation at long times is limited by the accumulation of errors inherent to approximate simulation. Here we introduce RL-Trotter, a reinforcement-learning framework that treats unavoidable approximation errors as resources for error correction rather than merely imperfections to suppress. We show that low-dimensional information from conservation laws, such as the energy and energy variance, provides a sufficient learning signal to guide the agent, which learns to adapt a single scalar---the next Trotter step size---without access to the target wave function. By optimizing the entire long-time evolution rather than individual steps, RL-Trotter discovers self-correcting sequences in which later errors compensate for those accumulated earlier, increasing the accuracy of the long-time dynamics. The learned policies are intrinsically robust to measurement noise, substantially reducing measurement overhead. They also generalize to previously unseen, physically similar initial states and transfer from small, classically simulable systems to systems an order of magnitude larger. This enables a practical protocol based on classical pretraining followed by direct deployment or limited fine-tuning on quantum hardware. Our results establish a broader perspective for quantum algorithms: errors in approximate evolution can be orchestrated into resources for accurate and resource-efficient quantum dynamics.", "5": "2026-08-22T19:13:20.438154"}
{"0": 170, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20110v1", "3": "ODEONN: A Digital ODE Solver Architecture for Oscillatory Neural Networks", "4": "ODEONN: A Digital ODE Solver Architecture for Oscillatory Neural Networks. Authors: Bram F. Haverkort, Aida Todri-Sanial. Categories: cs.AR. Oscillatory Neural Networks (ONNs) are an alternative computing paradigm for AI and combinatorial optimization problems. However, digital architectures are often designed for specific applications of ONNs. This work introduces a modular and scalable architecture called ODEONN that is generic to multiple applications of ONNs, and to the best of our knowledge, is the first fully digital ONN to also support complex-valued coupling. Additionally, an approximation of the sine function is introduced that uses half of the hardware resources compared to standard methods. The performance of ODEONN is compared with a full-precision software simulation, where a performance degradation of less than $2\\%$ is shown. Therefore, we conclude that the fixed-point quantization and the approximated waveform affect the accuracy of computation by only a small amount. Furthermore, ODEONN shows a 45$\\times$ reduction in energy-delay product over the software simulation running on conventional hardware.", "5": "2026-08-22T19:13:20.452426"}
{"0": 172, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20105v1", "3": "An AI-discovered smooth random fast dynamo on $\\mathbb{T}^3$", "4": "An AI-discovered smooth random fast dynamo on $\\mathbb{T}^3$. Authors: Keefer Rowan. Categories: math.AP math.PR. We construct a random, time-dependent divergence-free velocity field on $\\mathbb{T}^3$---refreshing iid on finite time blocks and obeying deterministic $C^\\infty_{t,x}$ bounds---that exhibits fast dynamo behavior. That is, for every fixed, sufficiently small resistivity, the almost sure exponential growth rate of the magnetic field solving the associated linear resistive induction equation is at least $1/2$; the exceptional null set may depend on the resistivity. We in fact get a time-uniform lower bound---with a random prefactor obeying a uniform-in-$\u03ba$ inverse moment bound. The argument relies on a particular algebraic structure in Fourier space of the induction equation solution operator that allows us to propagate expected growth of the logarithmic size of three specially chosen Fourier modes. This allows us to reduce to a simple recursion, avoiding the complicated infinite-dimensional dynamics typical to the dynamo problem. The central proof idea was generated autonomously by ChatGPT 5.6 Sol Ultra; the manuscript was written (and verified) by the author.", "5": "2026-08-22T19:13:20.461955"}
{"0": 174, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20053v1", "3": "On the Applicability of Safety Nets: A Safety-By-Design Solution for Certifying Neural Networks", "4": "On the Applicability of Safety Nets: A Safety-By-Design Solution for Certifying Neural Networks. Authors: Johann Maximilian Christensen, Thomas Stefani, Elena Hoemann, Frank K\u00f6ster, Sven Hallerbach. Categories: cs.AI. The integration of Artificial Intelligence (AI) in safety-critical aviation systems presents significant challenges for certification and deployment. Aviation, often regarded as the safest form of transportation, relies on numerous safety-critical systems. For future safety-critical AI-based systems, EASA requires a Safety-by-Design approach, which can be achieved by using Safety Nets that combine neural network compression with lookup tables to ensure 100 % correct runtime behavior across the discretized operational design domain. Although Safety Nets have been studied, no comprehensive study of their performance characteristics and system design trade-offs has been conducted. This work presents the first systematic analysis of the trade-off between neural network and lookup table size in Safety Nets. By systematically comparing neural networks with diverse architectures, this study identifies optimal design parameters that minimize overall storage and memory requirements while maintaining certification compliance. Results demonstrate that architectures with 3 to 5 hidden layers, each with approximately 50 to 100 nodes, combined with one-hot encoding, achieve the best balance. In these configurations, neural networks accurately represent at least 97 % of the data, while compact lookup tables handle the remaining errors. The resulting Safety Nets reduce the system size by almost three orders of magnitude, fitting within the memory budget of current avionics hardware while guaranteeing 100 % correct outputs across the entire discretized input space, as required by EASA guidelines. This work provides the first-ever open-source implementation of Safety Nets for HCAS and VCAS with replicable results, demonstrating a practical pathway toward certifiable AI-based systems in aviation and establishing Safety Nets as a viable Safety-by-Design solution for safety-critical applications.", "5": "2026-08-22T19:13:20.480524"}
{"0": 175, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20041v1", "3": "A three-dimensional typology of agency for advanced AI systems", "4": "A three-dimensional typology of agency for advanced AI systems. Authors: Willem Fourie. Categories: cs.AI cs.CY. Research on the agency of advanced artificial intelligence (AI) systems focuses on agency as a normative concept and on the agency of particularly agentic AI systems. While recent work also focuses on the different profiles of agentic systems, no framework exists to address the question of the type of agency instantiated by advanced AI systems, particularly when considering non-moral forms of agency. Based on established theoretical positions in philosophy, ethics, legal theory and sociology, we develop a typology of agency for frontier AI systems consisting of three dimensions: the nature of agency (moral or legal), its mode (individual or collective) and its locus (human or non-human). Combining these dimensions produces eight possible instantiations of agency, which we classify as conventional, contested or controversial. The typology separates legal from moral agency and thereby creates conceptual space for considering individual, legal, non-human agency without presupposing that advanced AI systems are moral agents. We argue that this distinction is increasingly relevant where instrumental goal pursuit complicates the attribution of AI actions to particular human actors.", "5": "2026-08-22T19:13:20.526466"}
{"0": 177, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20033v1", "3": "Temporal Fair Division of Indivisible Mixed Manna: Tractable Settings", "4": "Temporal Fair Division of Indivisible Mixed Manna: Tractable Settings. Authors: Kui-Wang Choi, Minming Li, Nicholas Teh. Categories: cs.GT. We study temporal fair division of indivisible mixed manna. Items arrive over time and must be allocated irrevocably; an item may be a good for some agents, a chore for others, and neutral for the rest. We require the cumulative allocation after every round to be envy-free up to one item (TEF1). Although deciding whether a TEF1 allocation exists is NP-hard even for goods, we identify several tractable settings. First, with at most $k$ item types, an online cyclic rule guarantees EF$\\lceil k/2\\rceil$ after every item arrival. Thus, every instance with at most two types admits an online TEF1 allocation; moreover, when the numbers of agents and types are fixed, TEF1 existence can be decided in polynomial time. Second, under agreement after agent-specific scaling, provided that the scaling factors are known before arrivals begin, an online rule produces an allocation that is EF1 and Pareto optimal after every item arrival. Third, for a two-part arrival sequence with common rankings, we give a rule that is EF1 after every item arrival. Fourth, when the number of agents is fixed and values are bounded integers, we give an exact pseudo-polynomial algorithm for deciding TEF1 existence. Finally, for goods, every TEF1 allocation gives each agent at least $1/n$ of her maximin share after every round; this factor is tight even for identical valuations and two rounds. Deciding whether an exact temporal maximin-share allocation exists is NP-hard for both goods and chores, even with identical valuations, two agents, and two rounds.", "5": "2026-08-22T19:13:20.545391"}
{"0": 178, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20026v1", "3": "From Street View Imagery to Street Quality Indicators: Vision Language Inference for the Suburban 15-minute City", "4": "From Street View Imagery to Street Quality Indicators: Vision Language Inference for the Suburban 15-minute City. Authors: Joan Perez, Giovanni Fusco. Categories: cs.CV cs.LG. Streetscape quality has become a central concern in contemporary urban planning, particularly within the framework of the pedestrian-friendly 15-minute city, where walkability and public-space quality are increasingly recognized as key determinants of urban performance. However, assessing streetscape qualities across large suburban and peri-urban territories remains challenging due to the time and resource demands of conventional field surveys. This paper presents a planning-oriented assessment of streetscape qualities in the north-eastern periphery of Nice (France) using the latest release of SAGAI (Streetscape Analysis with Generative AI), an open-source workflow that leverages vision-language models (VLMs) for large-scale streetscape analysis from Google Street View imagery. The new release addresses limitations of the original framework through improved image acquisition, geographically consistent view generation, support for multiple VLM architectures, consensus-based inference, and an integrated analytical environment. The workflow is applied to several thousand street-level observations to evaluate qualities relevant to pedestrian-friendly urban environments: sidewalk presence, pedestrian entrance density, and vegetation. The resulting maps reveal that the desired streetscape qualities characterize only a fraction of today's suburban streetscapes, mainly in compact developments and traditional suburban faubourgs, while they are particularly lacking on residential hills. The analysis demonstrates the potential of contemporary VLMs to support urban diagnostics in extensive suburban territories where fieldwork would be prohibitively time-consuming. Beyond the case study, the paper illustrates how recent advances in vision-language models can contribute to evidence-based planning by enabling scalable, flexible, and interpretable assessments of urban public-space quality.", "5": "2026-08-22T19:13:20.551552"}
{"0": 179, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20016v1", "3": "Emergence of cooperation: A reputation-modulated reinforcement learning", "4": "Emergence of cooperation: A reputation-modulated reinforcement learning. Authors: Chenyang Zhao, Jiqiang Zhang, Li Chen, Yong Zou. Categories: physics.soc-ph cond-mat.dis-nn nlin.AO q-bio.PE stat.ML. Reputation is widely recognized as a key mechanism for sustaining cooperation. However, most existing game-theoretic models treat reputation primarily as an external factor that modulates payoffs, interaction structures, or strategy update rules. In many social contexts, though, reputation operates primarily as information -- it shapes how individuals interpret their own experiences and assess the behavior of others. To bridge this gap, we propose a spatial prisoner's dilemma game grounded in the reinforcement learning paradigm, in which agents equipped with Q-learning integrate both individual and social information via a locally defined reputation metric to guide their decisions. Our results reveal that reputation-modulated learning significantly promotes the emergence of cooperative behavior, and we observe a discontinuous phase transition from full cooperation to full defection as the temptation increases. Cooperation spreads through the nucleation of cooperative clusters, whereas the disintegration of these clusters drives the system into an absorbing state of complete defection. Overall, this study demonstrates that reputation facilitates cooperation not only by providing direct incentives but also by reshaping the social information landscape that agents rely on for learning and adaptation.", "5": "2026-08-22T19:13:20.557888"}
{"0": 180, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19993v1", "3": "Optimal Skill Selection for LLM Agents with Provable Bicriteria Guarantees", "4": "Optimal Skill Selection for LLM Agents with Provable Bicriteria Guarantees. Authors: Yu Chen, Ruishuo Chen, Xun Wang, Zhuoran Li, Longbo Huang. Categories: cs.AI. Loading reusable skill documents into a bounded context window is now the primary way large language model (LLM) agents acquire task-specific capabilities, which makes skill selection a first-order determinant of task performance and token cost. Yet current agents score skills independently by semantic relevance and assemble the set by top-$k$ or greedy packing, with no quality guarantee or cost awareness on the selected set. As a result, redundant or poorly chosen skills waste scarce context tokens and can even degrade performance. We give the first model of how the selected skill set shapes execution outcomes and cast skill selection as an optimization problem: choose a skill set under a hard token budget to maximize a monotone submodular benefit minus context penalty. For this problem, we develop Best Prefix Selection (BPS), a polynomial-time algorithm, and prove, to our knowledge, the first performance guarantee for skill selection: a bicriteria $(1-1/e,1)$ approximation whose benefit coefficient is optimal in polynomial time. On a contamination-controlled BigCodeBench variant, BPS outperforms all the baselines, reaching $0.73$ measured task success versus $0.20$--$0.52$ for released skill routers, text retrievers, and the executor's own selection, on $28\\%$ fewer tokens than the strongest released router.", "5": "2026-08-22T19:13:20.561907"}
{"0": 192, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20305v1", "3": "CalcSeg: Confidence-aware 3D Latent Context Curriculum Learning For Myocardial Scar Segmentation From Single-Stack LGE-CMRs", "4": "CalcSeg: Confidence-aware 3D Latent Context Curriculum Learning For Myocardial Scar Segmentation From Single-Stack LGE-CMRs. Authors: Nivetha Jayakumar, Hannah Kim, Amit R. Patel, Miaomiao Zhang. Categories: cs.CV. Myocardial scar segmentation from single-stack late gadolinium-enhanced cardiac magnetic resonance (LGE-CMR) imaging has been a longstanding and clinically important challenge, particularly in the presence of low tissue contrast, diffuse, and small scar regions. These challenges are further intensified by the limited availability of 3D spatial context. This paper presents CalcSeg, a Confidence-aware latent context curriculum learning framework that leverages fused 3D feature representations from single-stack 2D LGE-CMR images for robust scar segmentation. Specifically, we introduce a dynamic semi-supervised curriculum learning strategy that progressively expands training from easier to more challenging scar cases using a learned confidence-aware scoring function. Such a function integrates errors in the predicted scar maps with quantified epistemic uncertainty and scar burden estimation to automatically assess sample difficulty without requiring manual labels. To compensate for the limited spatial context in single-stack acquisitions, we then develop a latent slice-wise self-attention to capture inter-slice dependencies and infer 3D spatial representations from sparse 2D inputs. We evaluate CalcSeg on multi-center clinical LGE-CMR datasets and benchmark against existing scar segmentation networks. Experimental results show that CalcSeg consistently outperforms all competing methods, particularly with substantial improvements on clinically challenging cases. Our code is released on Github.", "5": "2026-08-22T19:13:24.724711"}
{"0": 197, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20268v1", "3": "Learning Potts Models and $Z_3$ Toric Codes: Higher and Ordinary Nishimori Criticality", "4": "Learning Potts Models and $Z_3$ Toric Codes: Higher and Ordinary Nishimori Criticality. Authors: Rushikesh A. Patil, Malte P\u00fctz, Rohit Mukherjee, Guo-Yi Zhu, Simon Trebst. Categories: cond-mat.stat-mech cond-mat.dis-nn cond-mat.str-el quant-ph. Motivated by a previous Ising study, we identify a ${\\it higher}$ Nishimori line in the learning phase diagram of the $2D$ $q$-state Potts model $(2 < q\\leq 4)$ under bond-energy measurements. This ${\\it higher}$ Nishimori line meets the critical temperature line of the Potts model, in a ${\\it higher}$ Nishimori critical point -- a tricritical point at finite inference strength that separates a paramagnetic, a ferromagnetic and a 'spin-glass' phase. With analytical tools, we discuss the general structure of the rich phase diagram, which contains two unstable and three stable fixed points, and obtain a number of exact results for universal quantities, including the decay exponent of the Edwards-Anderson correlator, using a Gaussian measurement protocol which allows for exact calculations. Using extensive numerical tools, we confirm these statements for a generic, discrete $q$-state measurement protocol and determine precise numerical estimates for the location of higher and ordinary Nishimori critical points as well as RG flows between the various fixed points. We also discuss the Casimir effective central charges of the critical points in the learning phase diagram, and their monotonic ${\\it decrease}$ along measurement-induced RG flows, as established non-perturbatively by the c-effective theorem and its extensions, and contrast it to the monotonic increase along the corresponding RG flows in the random-bond Potts model. Finally, we discuss a general argument based on ${\\it Elitzur's \\; theorem}$ that establishes stability of the ordinary Nishimori critical points in their respective learning phase diagrams. Equivalently, our results describe a monitored deformed $\\mathbb{Z}_q$ toric code where the tricritical ${\\it higher}$ Nishimori point is an 'information' critical point that separates stable quantum, classical, and no memory phases.", "5": "2026-08-22T19:13:24.750308"}
{"0": 198, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20262v1", "3": "Probabilistic kilonova prediction from gravitational wave inferred binary neutron star parameters", "4": "Probabilistic kilonova prediction from gravitational wave inferred binary neutron star parameters. Authors: Xiao-Fei Dong, Ik Siong Heng, Gavin P. Lamb, Chris Messenger, Nikhil Sarin. Categories: astro-ph.HE gr-qc. Kilonovae provide a key electromagnetic window into binary neutron star mergers, revealing the properties of the merging system while probing r-process nucleosynthesis of the Universe. However, only a few kilonova candidates have been detected to date, with AT2017gfo being the only one associated with a gravitational wave event, GW170817. In this work, we present \\textsc{Genova}, a probabilistic framework for predicting kilonova spectra and light curves directly from gravitational wave posterior samples of binary neutron star mergers. This method uses a conditional normalising flow to learn the distribution of rest-frame spectra conditioned on the source-frame component masses, tidal deformabilities, viewing angle and time since merger. Other kilonova model parameters, such as ejecta opacities, are marginalised over during training, so that their effects are propagated into the predicted spectra as predictive uncertainty. In the self-consistency test, the flow model reproduces the median light curves with residuals typically below $\\sim 0.1$ mag, and the ratio of the predicted central 68\\% interval widths remains predominantly between $0.8$ and $1.4$ over $\\sim 0.4$--$8.0$ days. Comparisons with a physically distinct kilonova model show that the probabilistic prediction can remain informative beyond the model used for training. We apply \\textsc{Genova} to GW170817/AT2017gfo using multi-band observations, including newly re-reduced $Y$-, $J$-, $K_s$-band photometry from the Visible and Infrared Survey Telescope for Astronomy, which we present in this work. The resulting predictive intervals broadly encompass the observations while capturing both gravitational wave posterior uncertainty and the variation induced by marginalised kilonova model parameters.", "5": "2026-08-22T19:13:24.756900"}
{"0": 204, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20244v1", "3": "Extraction of DVCS amplitudes off the nucleon", "4": "Extraction of DVCS amplitudes off the nucleon. Authors: C. Mezrag, P. Sznajder, J. Wagner. Categories: hep-ph. This work presents a modern extraction of deeply virtual Compton scattering (DVCS) amplitudes off the proton and off the neutron through a global analysis of experimental data. These amplitudes serve as a crucial intermediate quantity linking generalized parton distributions (GPDs) to experimental observables. The analysis relies on a novel extraction framework recently integrated into the PARTONS software ecosystem. We employ two distinct modelling approaches for the DVCS amplitudes, both utilizing machine learning techniques: one model-agnostic and the other theory-augmented. For the first time, we incorporate the extraction of helicity-flip amplitudes, which are particularly sensitive to higher-twist effects. Furthermore, we compare our extracted amplitudes to timelike Compton scattering (TCS) data using the established relations between DVCS and TCS, serving as an important test of universality of GPDs. We also extract the DVCS subtraction constant, for which ab initio predictions from lattice QCD have been obtained. This allows us to confront these predictions with DVCS data for the first time, initiating a new virtuous cycle between first-principles theory and phenomenology.", "5": "2026-08-22T19:13:24.788849"}
{"0": 206, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20224v1", "3": "Intensity-based scattering correction enables in vivo two-photon imaging beyond 1 mm", "4": "Intensity-based scattering correction enables in vivo two-photon imaging beyond 1 mm. Authors: Yucheng Li, Renzhi He, Yi Xue. Categories: physics.optics. Optical imaging of the deep brain with subcellular resolution is essential for neuroscience, but noninvasive imaging beyond the cortex, through scattering white matter and into the hippocampus, has generally required three-photon microscopy at longer excitation wavelengths. Here, we introduce deep-learning-enhanced Fourier-domain intensity coupling for scattering correction (DeepFOCUS), an intensity-based two-photon approach that uses deep learning to compute intensity-modulation masks for real-time modulation of excitation light during image acquisition. Unlike deep-learning-based image restoration, this method directly improves image formation by computing intensity-modulation masks that shape the excitation light in real time, with each mask experimentally validated by the acquired fluorescence signal to avoid hallucination artifacts. Using 1035 nm excitation, we achieved in vivo two-photon imaging beyond 1 mm depth in the intact mouse brain, resolving YFP-labeled neurons and FITC-labeled blood vessels through the entire cortex and white matter down to the CA1 region of the hippocampus. DeepFOCUS extends two-photon imaging to depths previously accessible mainly with three-photon microscopy and could enable broader adoption of hippocampal imaging by upgrading existing two-photon systems.", "5": "2026-08-22T19:13:24.797895"}
{"0": 209, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20213v1", "3": "Explainability by Design: Structured Kolmogorov-Arnold Networks over Probabilistic Attributes for Speech Deepfake Source Tracing", "4": "Explainability by Design: Structured Kolmogorov-Arnold Networks over Probabilistic Attributes for Speech Deepfake Source Tracing. Authors: Hoang H. Pham, Manasi Chhibber, Tomi H. Kinnunen. Categories: eess.AS. Modern speech synthesizers can produce highly realistic speech, making source tracing (i.e. identifying the generator behind a spoofed utterance) increasingly important for forensics, online content provenance, and platform accountability. Building on our prior work on transparent probabilistic attributes, which represent utterances as probability distributions over synthesizer sub-components, we extend speech deepfake source tracing with two key ingredients: multi-task training of the probabilistic attribute extractors and a structured Kolmogorov--Arnold Network (KAN) for attack classification. The probabilistic features are estimated jointly with a multi-task learning module built on a shared AASIST or SSL-AASIST countermeasure backbone. The resulting probabilistic feature embedding is classified by a structured KAN whose topology follows known attribute-to-attack relationships. This provides interpretability by construction: the architecture reflects the generative hierarchy of attacks, while KAN feature-importance scores quantify each probabilistic feature's contribution without post-hoc explainers such as SHAP. On ASVspoof2019-attr-17, the extended framework achieves balanced accuracies above 99% for all seven probabilistic feature extractors, with EERs of 0.16% to 0.07%, and 99.64% balanced accuracy with 0.11% EER for 17-class attack classification. Our revised model outperforms the earlier two-stage baselines, in addition to demonstrating reliable interpretability, with importance scores consistent with SHAP values, and stable results across batch sizes. These findings highlight the potential of structured KAN for speech deepfake source tracing that is both accurate and interpretable by design. For transparency and reproducibility, our codebase is publicly available: https://github.com/HoangHPham/KAN-Probabilistic-Deepfake-Attribution.", "5": "2026-08-22T19:13:24.814462"}
{"0": 210, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20211v1", "3": "Inductive Process Discovery from Partially Ordered Event Data", "4": "Inductive Process Discovery from Partially Ordered Event Data. Authors: Humam Kourani, Tom Breuer, Gyunam Park, Wil M. P. van der Aalst. Categories: cs.DB. The Inductive Miner (IM) family is a prominent class of process discovery techniques, combining efficient recursive decomposition with soundness-by-construction guarantees. However, IM techniques usually assume traces to be totally ordered sequences of activity occurrences. This assumption is convenient, but can introduce systematic bias: activities may have durations, events may share coarse timestamps, or the data may constrain only some event pairs. Forcing such executions into arbitrary sequences hides inherent concurrency and may introduce sequential dependencies that were never observed as causal constraints. Partial orders provide a more faithful representation, but integrating them into IM discovery is challenging because standard abstractions are sequence-based; directly reusing them would require linearizing each partial order, which becomes prohibitively expensive under high concurrency. We introduce a lifting of IM discovery from total orders to partially ordered traces. Instead of redesigning the miner and its cut detection logic, we redefine the trace abstraction layer and the recursive projections to operate directly on partial orders. The approach is conservative over totally ordered traces, avoids linearization explosion, and preserves the recursive structure and guarantees that make IM attractive. Experimental results show that the proposed lifting avoids the combinatorial overhead of linearization, reduces sensitivity to arbitrary tie-breaking in timestamped event data, and allows process behavior to be learned from fewer observations by preserving concurrency at the trace level.", "5": "2026-08-22T19:13:24.819500"}
{"0": 228, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20111v1", "3": "Planning-Oriented End-to-End Autonomous Driving: Architectures, Evaluation, and Emerging Paradigms", "4": "Planning-Oriented End-to-End Autonomous Driving: Architectures, Evaluation, and Emerging Paradigms. Authors: Yanchen Guan, Xingcheng Liu, Bin Rao, Chengyue Wang, Guofa Li. Categories: cs.RO cs.ET. End-to-end autonomous driving has evolved from camera-to-control regression toward planning-oriented systems that use structured representations, trajectory-level outputs, and increasingly realistic evaluation protocols. This survey reviews this transition across behavior cloning, conditional imitation learning, privileged distillation, BEV and vectorized planning, unified perception-prediction-planning architectures, world-model-based planners, and vision-language-action systems. We argue that the key distinction in modern end-to-end driving is not whether intermediate representations are used, but whether they are learned, supervised, and evaluated to support safe, feasible, and route-compliant planning. To organize the literature, we synthesize existing methods along four axes: input representation, planning output, supervision signal, and evaluation protocol. We further examine the benchmark shift from open-loop trajectory matching to closed-loop simulation, non-reactive real-log evaluation, long-tail testing, and human-preference-aware metrics. Our analysis highlights that architectural progress is difficult to interpret without benchmark-consistent evaluation, and that displacement-based open-loop metrics alone provide limited evidence for safe and human-aligned driving. We conclude with open challenges in uncertainty-aware planning, learner-expert mismatch, runtime safety assurance, language-action grounding, world-model validation, and reproducible benchmarking.", "5": "2026-08-22T19:13:28.903464"}
{"0": 230, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20107v1", "3": "BeyondMasks: Evaluating Causal and Physical Consistency in Video Object Removal", "4": "BeyondMasks: Evaluating Causal and Physical Consistency in Video Object Removal. Authors: Yigit Ekin, Enes Sanli, Aykut Erdem, Erkut Erdem, Aysegul Dundar. Categories: cs.CV. Recent advances in generative video models have significantly improved visual realism in video object removal, yet evaluation protocols still focus on masked region fidelity, treating removal as local inpainting. In real scenes, object removal is a causal intervention: eliminating an object also requires removing its induced physical effects, such as shadows, reflections, illumination changes, translucency, and dynamic traces. Existing benchmarks lack aligned clean references or remain limited to simplified synthetic settings, preventing systematic evaluation of causal consistency. We introduce BeyondMasks, a paired benchmark for causally consistent video object removal, consisting of temporally aligned synthetic and real world video pairs with clean background references. The dataset spans diverse photometric, geometric, volumetric, and dynamic interactions, and supports both mask based and instruction driven editing. We further propose CORE, a structured vision language model based evaluation protocol that jointly measures object disappearance and after effect consistency, aligning more closely with human judgments than existing metrics. Benchmarking state of the art methods reveals systematic failures in removing secondary physical effects despite high masked region fidelity, exposing a gap between visual plausibility and causal correctness. BeyondMasks reframes video object removal as causal scene consistency rather than local reconstruction and provides a unified framework for its evaluation.", "5": "2026-08-22T19:13:28.913713"}
{"0": 238, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20011v1", "3": "Manifold Drift in Flow Preference Optimization: A Root Cause of Reward Hacking", "4": "Manifold Drift in Flow Preference Optimization: A Root Cause of Reward Hacking. Authors: Yansen Han, Shengyi Liao, Yuanxing Zhang, Pengfei Wan, Tao Lin. Categories: cs.AI cs.CV. Preference optimization is a standard alignment method for generative models, yet extending it to continuous-time dynamics remains non-trivial. In flow matching, reward-driven updates modify transport trajectories without an inherent constraint to the pretrained data manifold and can move terminal samples off the pretrained support. We formalize this failure mode as manifold drift. Theoretically, we show that optimal flow matching recovers the terminal data distribution, whereas a preference update leaves the pretrained manifold whenever its induced terminal displacement has a nonzero normal component. As a remedy, we propose ThermoDPO, a temperature-controlled objective that anchors pairwise preference optimization on preferred samples. Across temperature regimes, this objective connects rejection sampling fine-tuning and FlowDPO and controls a pointwise reconstruction-based surrogate for manifold distance. To counteract diminished signals at low temperatures, we further introduce a weighted variant, ThermoDPO-weighted. On the main toy benchmark, ThermoDPO-weighted attains a StrictScore of 0.899, compared with 0.629 for FlowDPO and 0.857 for FlowDPO+RFT. On SD3.5-M at CFG = 4.5, it improves OCR by 47.5% and the average of four metrics by 16.0%.", "5": "2026-08-22T19:13:28.955730"}
{"0": 239, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20007v1", "3": "Real Interference Alignment for Active IRS-Aided Systems: A Rate-Profile Learning-Based Approach", "4": "Real Interference Alignment for Active IRS-Aided Systems: A Rate-Profile Learning-Based Approach. Authors: Junda Liao, Quanzhong Li, Qi Zhang. Categories: cs.IT. With additional spatial degrees of freedom provided by the active intelligent reflecting surface (IRS), interference alignment (IA) can be achieved at low cost. In this letter, we propose a real IA scheme for an active IRS-aided system. The proposed scheme only requires the IRS to know the instantaneous channel coefficients under the assumption of blocked direct links. To maximize the achievable sum rate subject to individual minimum rate requirements and transmission power constraints, we propose a rate-profile learning-based algorithm. The algorithm uses offline-trained achievable rate profiles to decouple the original problem into multiple feasibility subproblems, which are then solved by generalized eigenvalue decomposition. Simulation results demonstrate that our proposed algorithm outperforms the conventional weighted minimum mean square error algorithm, while requiring significantly less program execution time.", "5": "2026-08-22T19:13:28.960313"}
{"0": 240, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20005v1", "3": "Scale-Aware Pretraining of Time Series Foundation Models via Multi-Patch Token Alignment and Hybrid Masking", "4": "Scale-Aware Pretraining of Time Series Foundation Models via Multi-Patch Token Alignment and Hybrid Masking. Authors: Taihua Chen, Xiang Ma, Yixin Zhang, Tailin Zhan, Manyu Sun. Categories: cs.LG. Pretraining time series foundation models across heterogeneous datasets necessitates effective handling of varying sampling frequencies. Current methods either employ dataset-specific patch sizes and separate FFNs, leading to fragmented representations, or enforce a fixed patch size that neglects inherent temporal variations. To address this, we propose SATS, featuring a scale-aware token alignment mechanism that treats patch size as an explicit notion of scale. By incorporating a contrastive-inspired alignment regularizer, SATS aligns representation spaces across scales while preserving distinct modeling capacities. Furthermore, a hybrid masking strategy combining random and contiguous masking is introduced to capture multi-scale temporal structures. Experimental results on LSTF benchmarks demonstrate that SATS achieves a 9.2% improvement in MSE and an 8.3% gain in GIFT-Eval MASE compared to competitive baselines. Notably, SATS consistently delivers SOTA performance while achieving a 65.6% increase in model efficiency over advanced baselines, highlighting its effectiveness and scalability in time series pretraining.", "5": "2026-08-22T19:13:28.968126"}
{"0": 249, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20174v1", "3": "Spin-Inversion Degeneracies in Restricted Inspiral Waveforms for LISA", "4": "Spin-Inversion Degeneracies in Restricted Inspiral Waveforms for LISA. Authors: Kata Kar\u00e1csonyi, L\u00e1szl\u00f3 \u00c1rp\u00e1d Gergely. Categories: gr-qc. Spin inversions appear in several settings. Analytical work predicts a single spin flip during unequal-mass supermassive binary inspiral, while numerical relativity and post-Newtonian calculations show repeated flip-flop motion in comparable-mass binaries. Secular spin evolution also predicts additional cases driven by spin-induced mass quadrupoles. Whether these effects can be distinguished in gravitational-wave data is still unclear. We combine the secular spin angle equations with a quasi-circular second post-Newtonian frequency evolution and build a restricted waveform weighted by the sky-averaged LISA sensitivity. We study five near-equal-mass injections with detector-frame total mass $M=2\\times10^5M_\\odot$, including Kerr flip-flops and one quadrupole-induced case. Each injection is compared with physically evolving waveforms constrained to have no orbital-plane crossings. We search all four no-inversion sectors, vary the masses, spin magnitudes, and initial spin angles, and maximize over time, phase, and overall amplitude. Large spin motion does not by itself lead to a clearly different waveform in this restricted model. A weak spin that sweeps through $133.26^\\circ$ leaves a residual SNR of $0.806$ at reference signal SNR $\u03c1_\\star=100$. A case where both spins cross the orbital plane many times gives the largest residual, $1.641$. The quadrupole case has a secondary-spin range of $128.11^\\circ$ with five crossings, yet the best no-inversion candidate found leaves a residual SNR of only $0.454$. All largest matches found exceed $0.999865$. Within this restricted model, the spin inversion cases are therefore strongly degenerate with no-inversion binaries at $\u03c1_\\star=100$. More complete waveforms, including observer-frame precession modulations, higher harmonics, separate polarizations, and the full LISA response, are needed to test whether this degeneracy can be broken.", "5": "2026-08-22T19:13:32.885714"}
{"0": 251, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20027v1", "3": "Electron energy gain in a dielectric laser accelerator as a function of the base angle of a triangular grating structure", "4": "Electron energy gain in a dielectric laser accelerator as a function of the base angle of a triangular grating structure. Authors: O. O. Svystunov, A. V. Vasyliev, I. V. Beznosenko, R. A. Melnichuk, G. V. Sotnikov. Categories: physics.acc-ph. Dielectric laser accelerators (DLAs) represent a compact and cost-effective alternative to conventional RF accelerators. Despite the various grating geometries already studied, a comprehensive investigation of triangular profiles, particularly the effect of the base angle of the saw-tooth grating profile, remains insufficient. This paper presents the results of numerical particle-in-cell (PIC) simulations of electron acceleration in DLAs based on double gratings with a triangular profile. One of the gratings, onto which the laser beam is incident, is transparent, while the second grating was either transparent or reflective for the laser pulse. The geometry of the first grating was fixed. A systematic study was conducted on the influence of the base angle of the second grating ($\u03b1= 5^\\circ - 44^\\circ$), its spatial orientation, the presence of a reflective gold coating, and the shape of the incident laser pulse (plane wave versus Gaussian profile) on the acceleration rate. The results reveal a complex interplay between these parameters. For an electron beam with an initial energy of 10 MeV, a maximum accelerating rate of 345 MeV/m was achieved for the structure with a left-handed reflective grating and a base angle of $\u03b1= 10^\\circ$ when excited by a Gaussian pulse. Under plane-wave excitation, the left-handed reflective grating with $\u03b1= 25^\\circ$ provided a rate of 325 MeV/m. It is shown that for single-bunch injection, nearly the maximum energy gain can be attained within a temporal window of 0.5 fs. The obtained results provide a quantitative basis for optimizing DLA designs with triangular gratings, highlighting the significant role of the saw-tooth profile base angle and the grating type in achieving high rates for next-generation accelerators.", "5": "2026-08-22T19:13:32.897758"}
{"0": 252, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20019v1", "3": "Contrastive Mixed Prompt Learning for Incomplete Multimodal Sentiment Analysis with Unseen Modality Combination", "4": "Contrastive Mixed Prompt Learning for Incomplete Multimodal Sentiment Analysis with Unseen Modality Combination. Authors: Kaixin Xu, NaiJin Liu, Yulin Kang, Tangyue Jin, Zixuan Yu. Categories: cs.AI. Incomplete multimodal sentiment analysis has garnered significant attention in recent years. Existing approaches typically assume that data is missing at random or are designed specifically for certain missing patterns, ignoring the modality combination inconsistency between training and testing phases. However, in real-world scenarios, the testing phase often encounters modal combinations that were not present during the training phase, which leads to insufficient generalization capabilities and unstable performance. In this paper, we introduce the problem of Incomplete Multimodal Sentiment Analysis with Unseen Modality Combinations (IMSAUMC), aiming to enhance model generalization for unseen modality combinations. To address this challenge, we propose the model named $\\textbf{C}$ontrastive $\\textbf{M}$ixed $\\textbf{P}$rompt $\\textbf{L}$earning ($\\textsf{CMPL}$) for IMSAUMC. It introduces a label-guided contrastive feature learning mechanism to learn robust and discriminative cross-modal representations. Additionally, we design modality-combination prompts with a soft router to facilitate better learning of various modality combinations. Furthermore, we introduce three prompt contrastive learning strategies, which enable effective learning of prompts corresponding to unseen modality combinations, thereby significantly strengthening the model's generalization capabilities in diverse testing scenarios. Extensive experiments on three widely used datasets demonstrate that $\\textsf{CMPL}$ achieves more than a 5% improvement in accuracy compared to state-of-the-art approaches.", "5": "2026-08-22T19:13:32.904206"}
{"0": 253, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19987v1", "3": "STEP: Score-Based Temporal Energy for Human Pose Video Anomaly Detection", "4": "STEP: Score-Based Temporal Energy for Human Pose Video Anomaly Detection. Authors: Jakub Micorek, Mateusz Kozi\u0144ski, Horst Possegger. Categories: cs.CV. Skeleton-based Video Anomaly Detection (VAD) offers a robust, privacy-preserving solution for identifying abnormal behaviors. To model the distribution of normal static and moving poses, recent methods train Energy-Based Models (EBMs) via Denoising Score Matching (DSM). However, directly injecting noise, required for training, into raw joint coordinates creates physically impossible poses, and this structural collapse severely worsens as the temporal window expands. To address this, we introduce STEP, a simple framework that utilizes Principal Component Analysis (PCA) to project pose sequences into a compact, whitened PC-space. Learning the data density within this well-behaved PC-space ensures that the injected noise translates into physically plausible variations, which allows the model to process longer video sequences without the performance collapse of raw coordinate baselines. Additionally, to mitigate inherent pose estimation inaccuracies arising from occlusions or motion blur, we integrate a sequence-level weighting mechanism based on the estimator's confidence scores. Operating at real-time computational efficiency, our simple and lightweight framework outperforms the previous skeleton-based state-of-the-art by 12.2% (90.1% AUROC) on the challenging UBnormal dataset and achieves highly competitive results by improving on the ShanghaiTech benchmark.", "5": "2026-08-22T19:13:32.910711"}
{"0": 255, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19974v1", "3": "ReguSim: Evaluating LLM Agent Rule Grounding in Financial Compliance", "4": "ReguSim: Evaluating LLM Agent Rule Grounding in Financial Compliance. Authors: Yiyang Luo, Yihang Jiang, Qijun Xie, Liang Lan, Lin Willian Cong. Categories: cs.AI. LLM agents in financial markets may cite rules yet still submit orders that violate executable constraints or misread surveillance evidence. We introduce ReguSim, a controlled financial-compliance environment, and ReguBench, a target-marked monitoring benchmark, to separate four artifacts: stated reasoning, attempted action, execution enforcement, and monitor evidence. In trader runs with DeepSeek V4 Pro and Gemini 3.5 Flash, visible rules reduce but do not eliminate rejected actions, and incentive or persona framing shifts behavior. A bridge study shows that trader rationales can mislead an independent monitor unless enforcement evidence is shown. In monitoring, simple structured baselines either match or exceed prompt-only LLMs. The results frame financial compliance evaluation as an audit of rule-grounded actions and evidence use, rather than a single compliance score.", "5": "2026-08-22T19:13:32.920293"}
{"0": 258, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19900v1", "3": "AvatarDynamizer: From Static to Dynamic Human Avatars via Generative Dynamic Textures", "4": "AvatarDynamizer: From Static to Dynamic Human Avatars via Generative Dynamic Textures. Authors: Guoxing Sun, Heming Zhu, Linjie Lyu, Pascal Fua, Christian Theobalt. Categories: cs.CV. For full-body avatars, modeling surface dynamics is crucial for overcoming the uncanny valley and achieving perceptual realism. Person-agnostic methods recover static 3D avatars from monocular images, videos, or text prompts, but their skeleton-driven animations lack realistic surface dynamics such as clothing wrinkles. In contrast, person-specific methods achieve high-quality rendering and realistic dynamics, but require expensive multi-view captures for each individual. Recent generalizable dynamic avatar methods struggle to embed surface dynamics, leading to either limited multi-view consistency or dynamic expressiveness. To this end, we propose AvatarDynamizer, a generative method that transforms an off-the-shelf static 3D avatar into a controllable, realistic, and multi-view-consistent 4D avatar. We introduce a novel texture-space surface-dynamics embedding and formulate avatar dynamics modeling as conditional texture generation. Our encoder--decoder representation embeds pose-dependent dynamics into dynamic texture maps, enabling compatibility with pre-trained video diffusion models while decoding them into 3D Gaussians for multi-view consistent rendering. Since existing datasets are limited in scale, sequence length, or motion diversity, we collect a large-scale multi-view dataset with long sequences covering diverse skeletal motions and surface dynamics. Experiments show that our method effectively animates static avatars with faithful surface dynamics and outperforms competing generalizable methods in visual fidelity, especially under limited dynamic training data.", "5": "2026-08-22T19:13:32.939003"}
{"0": 261, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19871v1", "3": "DIFFCZSL: Compositional Zero-Shot Learning Regularized by Diffusion Representations", "4": "DIFFCZSL: Compositional Zero-Shot Learning Regularized by Diffusion Representations. Authors: Hangyu Tian, Zhenqi He, Yanghao Wang, Long Chen. Categories: cs.CV. Compositional Zero-Shot Learning (CZSL) aims to recognize unseen attribute-object compositions by leveraging knowledge of primitive concepts learned from seen compositions. Although recent works achieve impressive performance in CZSL by leveraging large vision-language models, they primarily rely on discriminative representations that may not explicitly preserve the structured relationships between primitive concepts and their compositions. Motivated by the recent success of diffusion-based classifiers and their competitive performance relative to discriminative models, we investigate whether intermediate diffusion representations can provide complementary cues for CZSL. To this end, we propose DIFFCZSL, a diffusion-augmented framework that injects generative priors from pre-trained diffusion models into CLIP-based CZSL pipelines. We extract intermediate diffusion representations and project them into the CLIP embedding space to provide auxiliary supervision on both image and text modalities. Through contrastive alignment between CLIP embeddings and diffusion features during training, our method encourages the embedding geometry toward richer composition-aware semantics, while introducing no additional cost at inference time. Extensive experiments on three public CZSL benchmarks demonstrate consistent improvements over strong CLIP-based baselines under both closed-world and open-world settings. Our results highlight the complementary strengths of generative diffusion representations and discriminative vision-language models for compositional generalization.", "5": "2026-08-22T19:13:32.956238"}
{"0": 263, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19823v1", "3": "Simultaneous 3D co-registered perfusion and oxygenation with ULM, photoacoustic imaging, and a planar matrix array", "4": "Simultaneous 3D co-registered perfusion and oxygenation with ULM, photoacoustic imaging, and a planar matrix array. Authors: L\u00e9a Davenet, Jacques Battaglia, Franck Lager, Pascal Dargent, Charlotte Lussey-Lepoutre. Categories: physics.med-ph. Objective. Joint assessment of tissue oxygenation and microvascular perfusion could offer valuable insights into vascular function across a wide range of biomedical applications. Multispectral photoacoustic imaging enables the evaluation of blood oxygenation, while ultrasound localization microscopy provides sub-diffraction visualization of the microvasculature and blood perfusion. Here, we combine these two complementary modalities to simultaneously generate co-registered, volumetric maps of blood oxygenation and perfusion. Approach. Photoacoustic imaging and ultrasound localization microscopy are both ultrasound-based techniques. We developed an imaging platform that integrates the two modalities using a single planar ultrasonic matrix array, a state-of-the-art array for 3D ultrasound localization microscopy. The bimodal platform was validated in vitro using vessel-mimicking phantoms, then in vivo in mice. Main results. In vitro bimodal images of tubes injected with contrast agents demonstrated a coregistration accuracy of 20 $\u03bc$m and revealed complementary structural and functional information. Multispectral photoacoustic imaging achieved oxygen saturation measurements spanning the physiological range (60-95 %) with 5 % accuracy using only five optical wavelengths. In vivo imaging of healthy mouse tissues with known vascular anatomy further demonstrated the ability of the proposed platform to jointly characterize blood oxygenation and microvascular perfusion. Significance. This work experimentally validates a bimodal photoacoustic imaging-ultrasound localization microscopy approach using a planar ultrasound array. We characterized the functional imaging performance of this platform and identified limited-view artifacts inherent to this array configuration in photoacoustic imaging. These findings establish a foundation for adopting the platform in future studies of murine models and for advancing this promising bimodal approach.", "5": "2026-08-22T19:13:32.969581"}
{"0": 264, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19802v1", "3": "Stopping and Routing LLM Judge Panels", "4": "Stopping and Routing LLM Judge Panels. Authors: Bin Zhu, Yi Xie, Yanghui Rao. Categories: cs.CL. LLM evaluation pipelines often have many candidate judges: general LLM-as-a-judge prompts, reward models, safety classifiers, confidence variants, and task-specific verifiers. The deployment question is not only which judge is best, but which judges should be called, on which examples, and when panel construction should stop. We formulate judge-panel design as a role-conditioned allocation problem. From a small labeled audit set, declared slices, and judge costs, the method estimates target-relative roles: copies add no conditional information, complements improve the global panel, and specialists help only on slices. These roles induce a policy: drop copies, add complements globally, route specialists conditionally, and stop when validation gain falls below a threshold. Across reasoning, code, safety, preference, reward-model, summarization, and math audits, the method is compared with single judges, flat panels, matched diversity heuristics, full-call stacking, reliability juries, and frugal cascades. The result is a regime map for judge calls: route specialists on deployable slices, stop in saturated verifier regimes, keep broad ensembles when their risk benefit is worth the cost, and ignore conditional copies. The output is a reusable, auditable call plan for the next evaluation batch.", "5": "2026-08-22T19:13:32.975470"}
{"0": 265, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19748v1", "3": "Truncate Bad, Upweight Good: BoN-Style Distillation via Rank-Based Classification", "4": "Truncate Bad, Upweight Good: BoN-Style Distillation via Rank-Based Classification. Authors: Yarin Bar, Yaniv Romano. Categories: cs.LG cs.AI cs.CL. Inference-time selection methods, such as Best-of-N, improve generation by sampling a pool of candidates and selecting the top-ranked completion according to a reward model. Distillation seeks to amortize this procedure into a single policy by replacing raw rewards with in-pool ranks and learning a policy that upweights higher-ranked completions. However, existing rank-based policies typically use smooth full-support reweighting, so low-ranked completions receive less mass but remain in the target support. Although a sharper reweighting reduces lower-tail mass, it also increases reliance on brittle ranking at the top made by a single reward model. We propose TUP: a Truncate-bad, Upweight-good Policy that removes low-ranked completions from the support and reweights only the retained upper tail with a tunable sharpness. TUP admits a closed-form, prompt-independent normalization and can be trained fully offline via binary cross-entropy, using shifted-truncated win-rates as soft labels and distilled-to-reference log-likelihood ratios as logits. Theoretically, under certain assumptions, we show that for any unknown oracle reward, the best monotone rank-reweighting can be matched by a lower-tail truncation rule, providing formal support for removing the lower tail rather than merely downweighting it. Empirically, we show that TUP is competitive with strong offline alignment baselines.", "5": "2026-08-22T19:13:32.981038"}
{"0": 266, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19739v1", "3": "Question-Guided Evidence Acquisition for Multimodal Visual Question Answering", "4": "Question-Guided Evidence Acquisition for Multimodal Visual Question Answering. Authors: Alin-Ionut Popa. Categories: cs.CV cs.AI cs.LG. Multimodal LLMs can see a document, but they often can't read it reliably. Small text, tables, visual cues, and topological elements still trip them up under direct visual inference, even when the page is already sitting in the model's context. Most document-VQA systems treat perception as fixed: they encode the page once, ask the question, and answer from whatever the model happened to extract in that single fast pass. We think document VQA needs slower, more deliberate perception: rather than answering from one fixed encoding, the model should spend a bit of extra compute at inference time working out what to look at next, and only then answer. We build this into \\textbf{Q-Guide}, a small agent that reads a question, works out what evidence it is still missing, and calls targeted tool(s) to recover it---reading text where text is needed, zooming in where detail is needed, or grounding a region where position matters. On DocVQA2026 and Manga109, Q-Guide outperforms both direct prompting and recent multi-agent document systems ($65.0\\%$ vs.\\ $40.0\\%$ on DocVQA2026, $32.4\\%$ vs.\\ $24.4\\%$ on Manga109), and the improvement holds across three Claude backbones (Opus 4.6, Sonnet 4.6, and Opus 4.5). We find that accuracy scales with the perception budget---most of the gain appears within two to three deliberate rounds---and that the gain comes from directing perception to the right place, not from complex control logic: adding planners, routers, or multiple collaborating agents does not help.", "5": "2026-08-22T19:13:32.987414"}
{"0": 267, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19727v1", "3": "A Locally Tokenized Generative Model for Robust Time-Series Watermarking", "4": "A Locally Tokenized Generative Model for Robust Time-Series Watermarking. Authors: Dongbin Kim, Geonwoo Shin, Yujin Choi, Soyeon Park, Jaewook Lee. Categories: cs.LG cs.AI. Watermarking is a central tool for provenance in generative models, yet its application to multivariate time series remains hindered by reliability failures under post-editing attacks. We show that existing detectors, which rely on globally coupled re-encoding, suffer from bidirectional drift of the null distribution: post-editing attacks can shift the z-score of non-watermarked samples in either direction, invalidating clean-calibrated thresholds. We argue that this instability is a property of the re-encoding, and that reliable detection requires each recovered unit to depend only on a bounded temporal neighborhood. Guided by this principle, we propose L-VQVAE, a generative model in which each discrete token is produced from a short contiguous window, and LVQMark, a watermarking method over this token space that combines logit-bias injection with robust re-encoding for attack-time detection. Experiments on four benchmarks spanning finance, energy, and neuroimaging show that our approach preserves generation quality while stabilizing both detection power and false-positive behavior under post-editing attacks.", "5": "2026-08-22T19:13:32.991992"}
{"0": 268, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19714v1", "3": "Heavy-flavour production and correlations in pp collisions: precision tests of pQCD and hadronisation with ALICE", "4": "Heavy-flavour production and correlations in pp collisions: precision tests of pQCD and hadronisation with ALICE. Authors: Deependra Sharma, for the ALICE Collaboration. Categories: hep-ex nucl-ex. Heavy quarks (charm and beauty) are predominantly produced in hard partonic scatterings, making their cross sections in proton--proton (pp) collisions calculable in perturbative quantum chromodynamics (pQCD) and thus providing stringent tests of pQCD. Furthermore, the associated production of two charm hadrons in a single collision offers a sensitive probe of multiparton interaction dynamics, distinguishing between single parton scattering (SPS) and double parton scattering (DPS) processes.   In this contribution, preliminary measurements of prompt D-meson production are reported, together with the final results of the $\\rm B^{0}$-meson production cross section down to $p_{\\rm T}$ = 1 GeV/$c$ at midrapidity. The rapidity dependence of B-meson production is investigated by computing the ratio with respect to LHCb measurements at forward rapidity. The associated production of $\\rm D^{0}$--$\\mathrm{J}/\u03c8$ pairs in pp collisions at $\\sqrt{s} = 13.6$ TeV is presented as well, where $\\rm D^{0}$ mesons are reconstructed at midrapidity, while $\\mathrm{J}/\u03c8$ candidates are measured at forward rapidity. These measurements are compared with pQCD calculations and phenomenological models, providing crucial constraints on heavy-quark production, hadronisation, and multiparton interaction dynamics.", "5": "2026-08-22T19:13:32.999800"}
{"0": 269, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19665v1", "3": "Training-Free LLM-Based Recommendation with Post-LLM Item Refinement Using Collaborative Signals", "4": "Training-Free LLM-Based Recommendation with Post-LLM Item Refinement Using Collaborative Signals. Authors: Kyungho Kim, Sunwoo Kim, Geon Lee, Shinhwan Kang, Sojeong Kim. Categories: cs.IR. Large language models (LLMs) have shown promise for training-free recommendation, but LLM-generated user interests are often too broad for fine-grained item retrieval. Existing methods incorporate collaborative filtering (CF) signals in a pre-LLM manner through candidate reranking or prompt augmentation, yielding limited gains. We propose CoRRe, a training-free recommendation framework with a post-LLM paradigm that injects CF signals into LLM-generated item representations, which are later matched with LLM-generated user interests for ranking. Specifically, CoRRe refines the directions of item embeddings using an item-item co-purchase graph and their magnitudes using item popularity. Experiments on real-world datasets show that CoRRe consistently outperforms existing training-free methods and achieves competitive or superior performance compared with training-based methods, without requiring any model training or task-specific fine-tuning.", "5": "2026-08-22T19:13:33.004533"}
{"0": 270, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19644v1", "3": "When Guidance Goes Off-Scale: Recalibrating Diffusion Transformers under Analog Compute-in-Memory Nonidealities", "4": "When Guidance Goes Off-Scale: Recalibrating Diffusion Transformers under Analog Compute-in-Memory Nonidealities. Authors: Wenshuai Yao, Wenyong Zhou. Categories: cs.CV. Diffusion Transformers (DiTs) incur high memory traffic and energy costs because sampling repeatedly evaluates large denoisers dominated by linear operations. Analog compute-in-memory (CIM) can alleviate these costs by executing linear operations within weight-storing memory arrays. However, CIM nonidealities perturb effective weights, with errors accumulating along the state-dependent denoising trajectory; their interaction with classifier-free guidance (CFG) remains underexplored. In this paper, we characterize the impact of analog CIM nonidealities on DiT sampling. Although conditional and unconditional predictions can each remain close to their clean counterparts, their difference (the CFG residual) is disproportionately attenuated and rotated. Identifying this residual as a controllable failure channel, we propose a retraining-free, sampler-side recalibration that adjusts only the CFG scale for a given CIM condition. Trajectory-level analysis shows that moderate recalibration strengthens the target-oriented component preserved in the distorted residual, enabling earlier commitment to a prompt-consistent semantic region. In contrast, excessive guidance amplifies the full noisy residual and degrades quality, resulting in a finite, noise-dependent optimum. Extensive experiments on PixArt-Sigma, PixArt-alpha, and DiT-XL/2 show that the optimal guidance scale increases with CIM noise. Using 30,000 samples per condition, guidance recalibration consistently restores generation quality across simulated CIM mappings, closing at least 87% of the CIM-induced FID gap at a CIM noise level of 0.20. It reduces FID from 59.22 to 20.49 on PixArt-Sigma, 72.37 to 21.12 on PixArt-alpha, and 20.89 to 6.62 on DiT-XL/2.", "5": "2026-08-22T19:13:33.010928"}
{"0": 272, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20324v1", "3": "Correspondence between hydrodynamic frames, transport coefficients, and hydrodynamic modes in relativistic fluids", "4": "Correspondence between hydrodynamic frames, transport coefficients, and hydrodynamic modes in relativistic fluids. Authors: Md Hasanujjaman, Mahfuzur Rahaman. Categories: nucl-th hep-ph. The choice of hydrodynamic frame directly influences the numerical values of transport coefficients in relativistic dissipative hydrodynamics. We derive the exact transformation between the Eckart and Landau--Lifshitz frames and show that their thermal conductivities are related by an enthalpy-dependent factor. Using a baryon-rich relativistic fluid described by a Boltzmann nucleon gas equation of state, we find that the Landau--Lifshitz thermal conductivity is suppressed relative to the Eckart conductivity, with the difference increasing with temperature and baryon chemical potential. A linearized analysis of sound propagation demonstrates that the sound attenuation coefficient remains identical in both frames, confirming the frame invariance of physical observables. Our results show that the frame dependence of transport coefficients reflects only the different decomposition of dissipative effects into heat-flow and diffusion currents, while the underlying transport physics remains unchanged. These findings underscore the importance of specifying the hydrodynamic frame when comparing transport coefficients from heavy-ion collisions, lattice QCD, and kinetic-theory calculations.", "5": "2026-08-22T19:13:43.254093"}
{"0": 275, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20300v1", "3": "Taming the Tilt: A Unified Pilot Control Concept for Transformational eVTOL Aircraft", "4": "Taming the Tilt: A Unified Pilot Control Concept for Transformational eVTOL Aircraft. Authors: Daniel Milz, Marc May, Andreas Seefried, Tobias Bellmann. Categories: eess.SY math.OC. Transformational electric vertical take-off and landing (eVTOL) vehicles have gained significant attention over the past decade due to their efficient wing-borne cruise capabilities and reduced reliance on ground-based infrastructure. However, control system design for these vehicles remains challenging, as they must operate across multiple flight phases, each with distinct dominant dynamics. If left unaddressed, this complexity would significantly increase pilot workload, thus motivating the development of pilot control systems for multi-phase flight operations. The Simplified Vehicle Operations concept presents a promising strategy for reducing pilot workload. This study presents the design and implementation of a novel pilot control concept for eVTOL aircraft, validated through a tandem tilt-wing aircraft simulation on a full-motion simulator equipped with an active, force-feedback side stick. The system provides pilots with tactile feedback during specific flight phases, supporting intuitive control. The proposed approach enables seamless transitions and multi-phase flight maneuvers by leveraging the available degrees of freedom. Furthermore, an optimal-control-based methodology is proposed as a metric to evaluate command-filter-induced performance penalties and inceptor activities. The results show that the proposed command filter does not significantly increase the mission duration compared to the closed-loop system, while the active side stick helps reduce inceptor activity.", "5": "2026-08-22T19:13:43.269750"}
{"0": 276, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20291v1", "3": "Programmable cavity QED with a fiber-integrated atomic array", "4": "Programmable cavity QED with a fiber-integrated atomic array. Authors: Stephan Roschinski, Johannes Schabbauer, Franz von Silva-Tarouca, Marvin Holten, Damien Bloch. Categories: quant-ph cond-mat.quant-gas physics.atom-ph. Strong atom-photon interactions in optical cavities are a key resource for quantum information processing, quantum networking, and the exploration of quantum optical effects. Optical tweezer arrays offer scalable, site-resolved control of neutral atoms, but their integration with high-cooperativity cavity QED systems remains challenging. Here we combine a twelve-site $^{87}$Rb optical tweezer array with a high-cooperativity fiber Fabry-P\u00e9rot microcavity. The array is positioned within the cavity mode and individual sites are controlled with subwavelength precision, enabling continuous tuning of the single-atom coupling strength via deterministic displacement through the standing-wave field. For up to five atoms coupled to the cavity, we measure collectively enhanced vacuum Rabi splitting and implement cavity-based non-destructive readout of the number of coupled atoms. These results establish a scalable architecture for cavity-mediated entanglement generation and many-body cavity QED with single-atom control, and they lay the foundation for fiber-integrated quantum network nodes.", "5": "2026-08-22T19:13:43.274878"}
{"0": 279, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20267v1", "3": "Uniform weak type $(1,1)$ bounds for Riesz transforms on stratified Lie groups", "4": "Uniform weak type $(1,1)$ bounds for Riesz transforms on stratified Lie groups. Authors: Sheng-Chen Mao, Yaojun Wang, Ye Zhang. Categories: math.CA. Let $G$ be a stratified Lie group and $\\mathcal L$ its sub-Laplacian. We prove that the full horizontal Riesz transform $\\nabla_{H} \\mathcal L^{-1/2}$ is of weak type $(1,1)$ on real-valued functions, with constant at most $2$. In particular, the constant is independent of the horizontal dimension, the homogeneous dimension, the step, and the underlying group structure of $G$. Our result provides a noncommutative generalization of the dimension-free Euclidean theorem of Ouyang, Spector, and Stockdale arXiv:2608.18068 [math.CA], with the same universal constant. Our proof constitutes a fractional obstacle problem formulated through the heat semigroup and the associated Dirichlet form, adapted to the stratified structure. This extends the method of Ouyang, Spector, and Stockdale arXiv:2608.18068 [math.CA] without using the Fourier analytic ingredients.", "5": "2026-08-22T19:13:43.290524"}
{"0": 281, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20245v1", "3": "Coulomb blockade-like transport and enhanced memory in organic transistors embedded with sub-nm Pt nanoparticles for neuromorphic computing", "4": "Coulomb blockade-like transport and enhanced memory in organic transistors embedded with sub-nm Pt nanoparticles for neuromorphic computing. Authors: Arash Ghobadi, Thomas B. Kallaos, Abhi Abhijeet, Stephen C. Klue, Joseph C. Mathai. Categories: cond-mat.mes-hall. Organic transistors are playing an increasingly important role for neuromorphic applications. However, devices that rely solely on ferroelectric switching or on interface traps for their multi-conductance states exhibit limited memory windows. Here, we introduce an ultrathin oxide layer with a uniform distribution of sub-nm platinum nanoparticles (PtNPs) at the interface of a polymer semiconducting and a ferroelectric dielectric in a thin film transistor architecture. The interfacial stack, Al$_2$O$_3$/PtNP/Al$_2$O$_3$, provides a viable route for localized charge trapping and de-trapping in a region where it can most effectively influence the channel conductance. The organic transistors display a large memory window (> 20 V) in their current-voltage characteristics. The sub-nm PtNPs give rise to features that are consistent with room temperature Coulomb blockade-like transport, supporting discrete and well-separated levels within the memory window. The devices support multimodal programming using electrical and optical stimuli with both long-term plasticity and enhanced short-term plasticity (STP) phenomena. These results open new directions for implementing STP in the development of neuromorphic computing.", "5": "2026-08-22T19:13:43.301859"}
{"0": 284, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20227v1", "3": "Hybrid Qubit-Rotor Quantum Systems: Clifford Structure, Universal Control, and Applications", "4": "Hybrid Qubit-Rotor Quantum Systems: Clifford Structure, Universal Control, and Applications. Authors: Dengyao Luo, Arvin Kushwaha, Mastawal Tirfe, Bojko N. Bakalov. Categories: quant-ph. A $U(1)$ quantum rotor pairs a periodic angle with an integer-valued conjugate momentum, and occurs in molecular rotation, superconducting phase-charge circuits, and compact gauge fields. Coupling such a rotor coherently to qubits gives a hybrid register whose control structure is not inherited from either the oscillator-qubit or the qudit case. We develop a Clifford theory, together with a universal-control result, for registers of $n$ qubits and $r$ rotors. We classify all automorphisms of the hybrid phase space $\\mathbb{F}_2^{2n}\\times\\mathbb{Z}^r\\times\\mathbb{T}^r$ that preserve the Weyl commutation relations, and give an explicit finite Clifford circuit for each one. The classification is directional: rotor momentum parity may control qubit Pauli operations within the Clifford group, while every nonzero qubit-controlled rotor momentum shift is non-Clifford. It also yields normal forms for the mixed qubit-rotor couplings and the exact minimum number of elementary mixed gates needed to synthesize them. Adding a rotor cosine potential and one fixed qubit-rotor conditional phase to the local Clifford operations gives universal control on the full Hilbert space in the strong operator topology. We then apply this structure in three settings: an exact controlled-shift realization of gauge-covariant matter hopping, which is necessarily non-Clifford; rotor phase estimation with direct angle readout and probe optimization under momentum-support and energy constraints; and finite Fourier transforms on rotor momentum codes, where the one-rotor transform for $d=2^s$ compiles into $O(s)$ momentum-selective and controlled-shift instructions and each cross-register Fourier factor is implemented by one quadratic rotor Clifford gate.", "5": "2026-08-22T19:13:43.318071"}
{"0": 286, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20212v1", "3": "Unwarping the Lens: A Physics-Grounded Approach to Video Glasses Removal", "4": "Unwarping the Lens: A Physics-Grounded Approach to Video Glasses Removal. Authors: Radim Spetlik, David Futschik, Radek Danecek, Feitong Tan, Ziqian Bai. Categories: cs.CV. High-fidelity removal of eyeglasses from video is a major challenge in facial attribute editing, as the underlying facial geometry is often obscured by complex refractive distortions and view-dependent specular reflections. While large-scale generative priors have shown promise in eye-glasses removal via static image inpainting, they often lack the structural constraints necessary to maintain identity, expression, and pose, leading to visible \"identity drift\" in both static images and dynamic sequences. In this paper, we propose a novel transfer framework that addresses the stochastic nature of generative priors. Our pipeline first extracts high-fidelity synthetic face images from a commercial-grade generative model (Nano Banana, Gemini 3 Pro Image), regularizes them via a three-stage structural filtering process to preserve identity, expression, and pose, and finally applies physically-based simulation of lens optics during training to provide diverse, paired data. This process transfers Nano Banana's photo-realistic, multi-view knowledge into a specialized restoration architecture, JFSnet (Joint Feature-Spatial network). JFSnet integrates DINOv2-based semantic features with a convolutional decoder for spatial reconstruction, leveraging translation equivariance constraints to improve temporal consistency and high-frequency detail preservation. Evaluations on the curated Flickr-Faces-HQ (FFHQ) subset (12,163 images) show that our approach achieves high fidelity and structural accuracy, while maintaining inference speed of 27.68 FPS. In perceptual studies on CelebV-Text video sequences, our results are consistently preferred over diffusion and GAN-based baselines for ocular consistency, temporal stability, and overall restoration quality.", "5": "2026-08-22T19:13:43.327776"}
{"0": 289, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20198v1", "3": "A Resource-Efficient CNN-Based EEG Auditory Attention Decoding ASIC", "4": "A Resource-Efficient CNN-Based EEG Auditory Attention Decoding ASIC. Authors: Qier Ma, Richard George, Stefan Scholze, Jehn Constantin, Tobias Reichenbach. Categories: cs.AR eess.SP. Following a target speaker in a noisy environment, commonly known as the cocktail party problem, remains particularly challenging for cochlear implant (CI) users. Recent studies have explored EEG-based auditory attention decoding (AAD) using neural networks to enhance hearing assistance. This paper presents a resource-efficient ASIC for real-time EEG-based auditory attention decoding by integrating a quantized CNN inference engine and a Pearson-correlation classifier. The proposed architecture employs streaming execution, on-chip buffering, and memory-efficient dataflow to reduce hardware cost while maintaining real-time performance.   The proposed ASIC has been fully implemented in GF22FDX 22-nm CMOS technology, occupying a total silicon area of 2.09 mm$^2$(1264$\u03bc$m x 1654$\u03bc$m), with the CNN inference engine and streaming classification engine requiring only 0.076 mm$^2$. Operating at a core voltage of 0.55 V, the design achieves a power consumption of 0.4941 mW and an inference latency of 7.34 ms, providing an energy-efficient hardware platform for EEG-based auditory attention decoding in hearing-assistance applications.", "5": "2026-08-22T19:13:43.343447"}
{"0": 291, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20180v1", "3": "A Buchdahl solution generator for Einstein-sigma gravity and multiscalar-tensor gravity", "4": "A Buchdahl solution generator for Einstein-sigma gravity and multiscalar-tensor gravity. Authors: David S. Pereira, Francisco S. N. Lobo, Jos\u00e9 Pedro Mimoso. Categories: gr-qc. Buchdahl transformations provide a geometric route from vacuum spacetimes to scalar-field solutions. We formulate this mechanism as a target-covariant solution generator for Einstein gravity coupled to nonlinear sigma models in $D>3$ dimensions. On any connected region in which a Ricci-flat seed admits a non-null hypersurface-orthogonal cyclic coordinate with fixed-sign nonzero norm, the logarithm of that norm defines a Buchdahl potential $\u03a3$. A one-parameter redistribution of the cyclic norm then generates the exact Ricci source $R_{AB}=C_D(\u03b2)\\partial_A\u03a3\\partial_B\u03a3$, with $C_D(\u03b2)=\\frac{D-2}{D-3}(1-\u03b2^2)$, while $\u03a3$ remains harmonic in the deformed metric. Composing $\u03a3$ with an affinely parametrized geodesic of a non-degenerate target metric whose squared speed is $C_D(\u03b2)$ therefore yields an exact Einstein--sigma-model solution. For positive-definite targets, the corresponding pullback condition also implies that the nontrivial scalar map has one-dimensional image; for indefinite targets, the geodesic ansatz remains a sufficient solution-generating sector but need not exhaust all possible realizations of the same rank-one spacetime source. Through the Einstein/Jordan-frame correspondence, the construction gives vacuum tensor--multiscalar solutions as well. We illustrate the unified generator with flat, spherical, hyperbolic and axion--dilaton targets and with Tangherlini, Weyl, Kasner, Rosen-wave, Gowdy, Kaluza--Klein bubble and C-metric seeds. The result separates cleanly the geometric Buchdahl source from its target-space realization and extends this rank-one cyclic sector across static, cosmological, wave and accelerating geometries.", "5": "2026-08-22T19:13:43.354113"}
{"0": 292, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20165v1", "3": "Orbit equivalence and total weak mixing of free group actions", "4": "Orbit equivalence and total weak mixing of free group actions. Authors: Konrad Wr\u00f3bel. Categories: math.DS math.GR math.LO. We prove that the orbit equivalence class of every free ergodic probability-measure-preserving (pmp) action of a free group contains a totally weak mixing action. Equivalently, every ergodic treeable pmp equivalence relation of cost $n\\in\\mathbf{N}\\cup\\{\\infty\\}$ is generated by a free totally weak mixing action of $\\mathbf{F}_n$. This answers a question of Miller and Tserunyan.   The proof goes by considering a Polish space of edge slidings along a fixed mixing transformation and proving that for every $w\\not=e\\in\\mathbf{F}_n$ the set of edge slidings that produce an action with $w$ weakly mixing forms a comeager set.", "5": "2026-08-22T19:13:43.362179"}
{"0": 293, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20164v1", "3": "Architecture and Compilation Co-Design for High-Rate Quantum Product Codes on Neutral Atom Arrays", "4": "Architecture and Compilation Co-Design for High-Rate Quantum Product Codes on Neutral Atom Arrays. Authors: Adrian Liu, Wan-Hsuan Lin, Daniel Bochen Tan, Qian Xu, Jason Cong. Categories: quant-ph cs.AR. Achieving fault-tolerant quantum computing at a practical scale demands quantum error correction (QEC) codes with high encoding rates. Quantum low-density parity-check (qLDPC) codes emerge as a promising candidate, especially given the rise of neutral atom arrays that provide dynamic long-range connectivity via atom movements. In general, synthesizing valid and efficient physical execution plans for QEC is a provably hard combinatorial problem, forming a critical compilation bottleneck that worsens as code sizes grow. To overcome this complexity, we focus on an important product family of qLDPC codes with dimension-reduction properties, and propose ONEX. This framework decomposes complex 2D physical execution planning into independent 1D subproblems, each solved to optimal execution depth within practical compilation time. First, we formulate the 1D execution plan with an explicit satisfiability modulo theories (SMT) encoding. This protocol produces provably depth-optimal solutions with substantial duration reduction. Second, we develop a multi-stage compilation pipeline featuring anytime optimization, movement compaction, and iterative feedback. This pipeline maintains practical wall-clock times while providing progressive refinement and on-demand retrieval of quality solutions. Third, we evaluate ONEX in the application of hypergraph product (HGP) code memory mapped onto neutral atom arrays, achieving 3.7x to 6.1x and 29.8x to 42.1x higher clock rates than the constructive 1D algorithm and the general 2D compiler, respectively, while scaling efficiently to codes with 2,500 data qubits. Finally, we extend ONEX to zoned layouts, revealing architectural insights into the associated trade-offs, and demonstrate its applicability to the broader lifted-product (LP) code family through a representative example.", "5": "2026-08-22T19:13:43.366829"}
{"0": 295, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20148v1", "3": "Using Zone-Disjoint Multi-Path Routing Algorithm for Video Transmission over Ah-Hoc Networks", "4": "Using Zone-Disjoint Multi-Path Routing Algorithm for Video Transmission over Ah-Hoc Networks. Authors: Ali Asghar Khavasi, Nastooh Taheri Javan. Categories: cs.NI. Finding multi-path routes in ad hoc networks due to their grid topology seems to be a trivial task, but because of CSMA/CA effects in these networks found paths are not completely disjoint unless an appropriate algorithm have taken into account. If such an algorithm provided and designed carefully it could improve multi-path video transmission over these kinds of networks. By using node-disjoint paths, it is expected that the end-to-end delay and BER in each case should be independent of each other. However, because of natural properties and medium access mechanisms in ad hoc networks the end-to-end delay and also BER between any source and destination depends on the pattern of communication in the neighborhood region. In this case some of the intermediate nodes should be silent to reverence their neighbors and this matter increases the end-to-end delay. To avoid this problem, multi-path routing algorithms can use zone-disjoint paths instead of node-disjoint paths. In this paper we demonstrated a new multi-path routing algorithm that selects zone-disjoint paths that in addition used for video multi-path transmission over ad hoc networks. It is shown that by using this new algorithm along with choosing an appropriate scheme for video transmission over discovered paths, our approach receives enhanced results comparing previously used algorithms.", "5": "2026-08-22T19:13:43.379656"}
{"0": 296, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20144v1", "3": "PelviNeXt: A Modality-Agnostic Hybrid Network for Pelvic Imaging in Women's Health", "4": "PelviNeXt: A Modality-Agnostic Hybrid Network for Pelvic Imaging in Women's Health. Authors: Siam Tahsin Bhuiyan, Rashedur Rahman, Sefatul Wasi, Halima Khatun, Ashraful Islam. Categories: cs.CV. Women's health remains substantially under-resourced in medical imaging research, with pelvic pathologies such as polycystic ovary syndrome (PCOS) and pelvic fracture both suffering from a scarcity of public, well-annotated benchmark data despite their clinical importance. We introduce PelviNeXt, a modality-agnostic hybrid architecture combining a dense convolutional feature extractor, hierarchical channel-spatial attention (H-CBAM), a multi-scale fusion module (MSFM), and talking-heads multi-head self-attention (TH-MHSA), applied without modification to both pelvic ultrasound and X-ray inputs. While benchmarking PelviNeXt on PCOSGen, the only gynaecologist-annotated public PCOS ultrasound dataset, we identified extensive exact and near-duplicate contamination within and across the dataset. We audit this contamination via perceptual hashing, publicly release a deduplicated version of the dataset, and establish the first integrity-audited evaluation protocol and baseline for PCOSGen under 5-fold cross-validation. On the only publicly available pelvic fracture X-ray dataset (PXR150), PelviNeXt exceeds previously reported state-of-the-art results across accuracy, recall, specificity, and AUROC. Ablation studies confirm that each architectural component contributes to performance on both tasks. Our results demonstrate that a single architecture, applied without task-specific modification, can serve as a reliable foundation for pelvic imaging across modalities in data-scarce, under-researched areas of women's health.", "5": "2026-08-22T19:13:43.385389"}
{"0": 297, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20137v1", "3": "Formal Performance and Compile Time Guarantees for Compiler Optimization Heuristics", "4": "Formal Performance and Compile Time Guarantees for Compiler Optimization Heuristics. Authors: Nikil V. Shyamsunder. Categories: cs.PL cs.PF. Modern optimizing compilers rely on heuristic search algorithms for NP-hard optimization problems, which can result in poor generated-code performance and long or unpredictable compile times. These are considered bugs by users, but verified compilers rarely reason beyond semantic preservation. We propose verifying performance and compile time properties of compiler passes. As a proof-of-concept, we formulate inline expansion using a cost model estimating instruction-cache performance. We mechanize this in Rocq, prove semantic preservation of the inlining transformation, and verify the algorithm's monotone improvement, convergence-time bound, and performance bounds for intermediate and final solutions.", "5": "2026-08-22T19:13:43.391925"}
{"0": 299, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20127v1", "3": "ID-VTG: Image-Disambiguated Video Temporal Grounding", "4": "ID-VTG: Image-Disambiguated Video Temporal Grounding. Authors: Minghang Zheng, Jingli Wei, Hongyi Yang, Yang Liu. Categories: cs.CV. Video Temporal Grounding (VTG) faces significant challenges when natural language queries must distinguish between multiple events involving visually similar entities, particularly when relying on fine-grained visual attributes that are difficult to describe accurately in words alone. To address this, we introduce Image-Disambiguated Video Temporal Grounding (ID-VTG), a task that leverages multimodal queries combining a reference image and a text description to precisely localize segments where a specific instance performs a described action. To facilitate research, we construct two benchmarks: IDVTG-Gym, focusing on fine-grained, compositionally ordered gymnastics actions with athletes in similar uniforms; and IDVTG-InternVid, an open-world dataset featuring diverse entities (e.g., humans, animals, fictional characters) and significant temporal distractors. Methodologically, we propose the Visually-Guided Disambiguation Aggregation (VGD-Agg) framework based on a dual-branch fast-slow architecture. The fast branch efficiently generates preliminary event proposals, while the slow branch performs fine-grained frame-level matching between video frames and the reference image. We enhance discriminability via two learnable tokens: a Compare Token, which represents hard negatives to probe for the presence of the target instance (as referred to by the query image), and a Depress Value, which represents text-irrelevant events. Proposals that the Compare Token identifies as lacking the target instance are pushed toward the Depress Value, thus easing disambiguation via the text query. Extensive experiments validate our approach, which achieves state-of-the-art results on the proposed benchmarks. Code is available at https://github.com/oceanflowlab/ID-VTG.", "5": "2026-08-22T19:13:43.401635"}
{"0": 300, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20122v1", "3": "ArmorOCR: Grounded Adversarial Visual Perception via Observation-Transferred Self-Distillation", "4": "ArmorOCR: Grounded Adversarial Visual Perception via Observation-Transferred Self-Distillation. Authors: Linhan Cao, Siyuan Li, Jun Lan, Liangbo He, Guannan Li. Categories: cs.CV. Large multimodal models (LMMs) have demonstrated strong OCR recognition capabilities, yet remain vulnerable to adversarial visual text that is readable to humans but challenging for models to localize and recognize. Existing OCR benchmarks mainly focus on natural or document-style text, while adversarial OCR evaluations remain limited in scale, task coverage, or region-aware evaluation. In this paper, we formulate adversarial OCR as a \\textbf{grounded OCR perception} task and introduce \\textbf{AdvSpot}, the first benchmark for grounded adversarial OCR evaluation. AdvSpot comprises 390 images with region-level annotations, spanning 5 primary categories and 13 fine-grained adversarial OCR types. To address this challenge, we propose \\textbf{ArmorOCR}, a two-stage training framework for robust adversarial OCR perception. ArmorOCR first acquires missing adversarial OCR perception from privileged transformed observations through On-Policy Self-Distillation (OPSD), and then refines grounded OCR perception through Group Relative Policy Optimization (GRPO) with task-conditioned rewards for localization, recognition, full spotting, and visual question answering (VQA). Experiments on our AdvSpot, other adversarial OCR benchmarks, and general OCR benchmarks demonstrate that ArmorOCR consistently improves adversarial OCR perception while preserving competitive general OCR capability.", "5": "2026-08-22T19:13:43.407355"}
{"0": 342, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20293v1", "3": "Non-Minimally Coupled Warm Inflation in the Defining Frame", "4": "Non-Minimally Coupled Warm Inflation in the Defining Frame. Authors: Adri\u00e1n Casado-Turri\u00f3n, Paulo B. Ferraz, Mindaugas Kar\u010diauskas, Jos\u00e9 Jaime Terente D\u00edaz. Categories: astro-ph.CO gr-qc hep-ph. Warm inflation in $F(\u03a6)R$ scalar-tensor theories of gravity is investigated in the `defining' frame, where the theory and its parameter values are specified. Translating the resulting dynamics to the Einstein frame, we find that the dissipation ratio is suppressed by the modified-gravity effects. Thus, although the effective warm-inflation dynamics can be consistently analysed in either frame, the dissipative regimes need not coincide between them. In particular, we find that quantum perturbations can dominate over thermal fluctuations in the scalar power spectrum even in a high-temperature, strong-dissipation regime in the defining frame. Finally, we compute the scalar spectral index and tensor-to-scalar ratio for a non-minimal coupling function $F(\u03a6) = 1+\u03be(\u03a6/m_{\\rm P})^2$ with a quartic potential and both constant and quadratic field-dependent dissipation coefficients, and identify benchmark points compatible with current CMB constraints.", "5": "2026-08-22T19:13:52.035200"}
{"0": 353, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20208v1", "3": "RoMAN-Flow: Taming Autoregressive Normalizing Flows for Offline Reinforcement Learning in Robotic Manipulation", "4": "RoMAN-Flow: Taming Autoregressive Normalizing Flows for Offline Reinforcement Learning in Robotic Manipulation. Authors: Shaoxuan Wang, Guangting Zheng, Rui Huang, Zhipeng Tang, Sha Zhang. Categories: cs.CV. Offline reinforcement learning improves robotic policies using previously collected data without further environment interaction. Yet prevalent diffusion- and flow-matching robot policies lack tractable likelihoods, limiting their use in likelihood-based offline RL post-training. AR-NFs offer both expressive action modeling and exact likelihood evaluation, but their sequential sampling incurs substantial sampling overhead during policy optimization and deployment. We present RoMAN-Flow (Robotic Manipulation with Autoregressive Normalizing Flows), an offline reinforcement learning framework that makes AR-NF policies practical for robotic manipulation by addressing this sampling bottleneck in both stages. During policy optimization, RoMAN-Flow employs a sampling-free, advantage-weighted likelihood objective that assigns higher likelihood to high-advantage actions from the offline dataset without sampling from the autoregressive policy. For efficient deployment, it distills the optimized autoregressive policy into a one-step action generator, enabling low-latency action prediction. Experiments across multiple simulated manipulation benchmarks and real-world robotic platforms demonstrate that RoMAN-Flow achieves competitive policy performance while substantially reducing inference latency. Code is available at https://github.com/konnyaku28/RoMAN-Flow.", "5": "2026-08-22T19:13:52.090518"}
{"0": 356, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20193v1", "3": "The 2PN Point-Mass N-Body Equations of Motion in Harmonic Gauge: A Computable Formulation", "4": "The 2PN Point-Mass N-Body Equations of Motion in Harmonic Gauge: A Computable Formulation. Authors: Hongkun Huang, Jie Yang, Wei-Tou Ni. Categories: gr-qc. We develop a semi-analytic and semi-numerical formulation of the harmonic-gauge second post-Newtonian (2PN) equations of motion for a general point-mass N-body system within Hadamard regularization. The equations of motion are separated into a closed analytic contribution and a non-closed integral contribution. We analyze the singular structure of the latter and further regularize it into a numerically evaluable representation. We apply the formulation to the Sun-Jupiter-Saturn and Sun-Mercury-Venus systems, evaluating the instantaneous non-closed 2PN acceleration along Newtonian trajectories and its leading finite-time relative-distance response through the corresponding perturbation equations. In both benchmarks, the non-closed acceleration remains a small fraction of the complete 2PN acceleration, while the induced relative-distance perturbation remains oscillatory and can reach larger oscillation amplitudes at later times.", "5": "2026-08-22T19:13:52.105135"}
{"0": 358, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20185v1", "3": "An Analytical Two Incompressible Fluid Star with a Mixed Ordinary Dark Matter Core and an Ordinary Matter Envelope", "4": "An Analytical Two Incompressible Fluid Star with a Mixed Ordinary Dark Matter Core and an Ordinary Matter Envelope. Authors: Milko Estrada, Santiago Esteban Perez Bergliaffa. Categories: gr-qc. We construct an analytical relativistic two fluid star characterized by a mixed core, where ordinary matter and dark matter coexist as two independently conserved incompressible perfect fluids, and an envelope composed exclusively of ordinary matter. The fluids exchange neither matter nor energy and interact only through the common spacetime geometry, with the ordinary component extending across the core envelope interface while the dark component is confined to the core. Despite the mixed core--single-fluid envelope structure and the internal interface, the system remains analytically tractable, allowing us to obtain explicit expressions for the pressures and metric functions and to follow directly the effects of the dark-matter fraction and relative core size. We determine the physically admissible parameter space and derive a Buchdahl like critical compactness associated with the divergence of the central pressure, whose value depends on the relative dark matter density and the size of the mixed core. The Schwarzschild constant density star and its standard critical value, $2M/R=8/9$, are recovered in the corresponding one fluid limit. The mass--radius analysis further shows that configurations with the same global compactness can correspond to distinct internal matter distributions. Beyond providing an analytically controlled description of a core-confined second component, the construction offers a useful benchmark for identifying qualitative trends that may subsequently be examined in more realistic dark matter admixed neutron star models, whose detailed treatment lies beyond the scope of the present work.", "5": "2026-08-22T19:13:52.118578"}
{"0": 359, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20184v1", "3": "PEtab SciML: an exchange format for specifying and training dynamic scientific machine learning models", "4": "PEtab SciML: an exchange format for specifying and training dynamic scientific machine learning models. Authors: Sebastian Persson, Branwen Snelling, Maren Philipps, Daniel Weindl, Marija Cvijovic. Categories: q-bio.QM. Summary: Dynamic scientific machine learning (SciML) models that combine mechanistic ordinary differential equations (ODEs) with machine learning (ML) components have applications ranging from learning unknown biological processes to integrating auxiliary data modalities into dynamic modelling. To enable reproducible and efficient SciML training, we introduce PEtab SciML, an interoperable data format for specifying parameter estimation problems in which mechanistic and ML model parameters are jointly estimated from time series data. PEtab SciML supports several ML ODE hybridization patterns in realistic problem setups. It is accompanied by a reference Python library and downstream modelling support in Python/JAX and Julia, provided by AMICI and PEtab.jl, respectively, and a collection of real data benchmarks. Availability and implementation: PEtab SciML is available on GitHub (https://github.com/PEtab-dev/petab_sciml). The reference Python package is installable from PyPI and is continuously tested and supported on Linux, macOS, and Windows.", "5": "2026-08-22T19:13:52.124182"}
{"0": 428, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20304v1", "3": "Calibration-Induced Degeneracy in LLM Financial Forecasting: An Audit-Trailed Case Study on Next-Day Market Risk", "4": "Calibration-Induced Degeneracy in LLM Financial Forecasting: An Audit-Trailed Case Study on Next-Day Market Risk. Authors: Arin Mohanty. Categories: q-fin.RM. Costly LLM features matter only if calibration lets them affect the forecast. We document a failure of this link in a next-day risk study of two broad-market funds. Full-history scoring preceded the 2022 calibration. Calibration then set all four LLM weights to zero. The 856 later scores therefore could not affect the evaluation. We call this calibration-induced degeneracy. Allowing signed weights reactivated all four mappings. None improved forecasts after familywise correction. By contrast, a near-zero-cost headline count reduced SPY variance-forecast loss by 0.001720 (95 percent familywise interval: [0.000719, 0.002830]). The cheap baseline is therefore a critical diagnostic. We propose a calibration-viability checkpoint. Fit the mapping, perturb the feature over prespecified calibration values, and require a meaningful forecast response before acquiring holdout features. The check uses no holdout outcomes. Here, it would have stopped the paid full-history inference phase.", "5": "2026-08-22T19:14:05.392761"}
{"0": 430, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20279v1", "3": "Robustness of random-walk Metropolis for steep potentials", "4": "Robustness of random-walk Metropolis for steep potentials. Authors: Sam Power. Categories: math.PR math.ST stat.CO. In Markov chain Monte Carlo sampling, light-tailed target distributions present something of a poisoned chalice: their light tails offer good confinement, and tend to imply good mixing properties for natural continuous-time dynamics, but the steepness of their tail decay means that they often fall outside of the scope of modern quantitative convergence theory. For usual gradient-based samplers, this reflects a genuine instability issue, whereby Metropolis acceptance rates can degrade badly. In this work, we study the gradient-free random-walk Metropolis sampler, and show that for a wide range of light-tailed targets, the acceptance probability remains stable for reasonable choices of proposal variance, from which effective and favourable mixing time estimates can be deduced. The analysis relies on a simple relationship between the first and second derivatives of the log-density of the target distribution.", "5": "2026-08-22T19:14:05.403372"}
{"0": 439, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20167v1", "3": "BreakGuard: Towards Detecting Dependency Breaking Changes with LLM-Generated Tests", "4": "BreakGuard: Towards Detecting Dependency Breaking Changes with LLM-Generated Tests. Authors: Rachna Raj, Benoit Baudry, Diego Elias Costa. Categories: cs.SE. Open-source libraries play an important role in software development by providing reusable features that expedite the development process. As libraries evolve, they release new versions that add features, fix bugs, or apply security patches. In this process, they may break the contract established with their clients by introducing breaking changes (BCs) that alter the runtime behavior and break client applications. Client-side test suites often fail to detect these BCs because of limited library coverage that does not exercise all library methods used in the client's codebase.   We propose BreakGuard, an approach that generates a test suite to detect breaking changes in clients. BreakGuard statically extracts every client method (focal method) that invokes the target library method (call site), then generates tests per focal method. A test detects a BC if it passes on the pre-breaking version and fails on the breaking version. We evaluate our approach on 89 real-world breaking changes from the BUMP dataset, using 3 LLMs (GPT4o, Qwen3-coder-480B, GPT-OSS-120B) and three context levels: minimal, method, and class. Using the best-performing configuration, BreakGuard detects 30.3% of breaking changes (27 of 89) at a mean cost of roughly $0.90 USD per detected breaking change. We successfully detected BCs from different library categories (e.g., JSON libraries, logging, parsing), but we find LLM-generated tests to be more reliable for detecting crash-type breaking changes as opposed to behavioural BCs.", "5": "2026-08-22T19:14:05.448613"}
{"0": 448, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20086v1", "3": "Correlations at criticality in ecological communities", "4": "Correlations at criticality in ecological communities. Authors: Akiva Goldberg, Nadav M. Shnerb. Categories: q-bio.PE. Ecological communities are continually reshaped by invasion, exclusion, and diversification, processes that naturally drive them toward the boundary of dynamical stability. Near such a boundary, a soft mode relaxes increasingly slowly and, under stochastic forcing, is expected to dominate the fluctuations, effectively reducing the dynamics to one dimension and generating strong positive and negative abundance correlations. Such correlations have therefore been proposed as signatures of an imminent transition. Here we show that this expectation can fail even arbitrarily close to criticality. The reason is that spectral softness does not guarantee stochastic visibility: the soft mode must receive enough environmental forcing to dominate the fluctuation background generated by the remaining modes. We demonstrate this mechanism in three ecological scenarios: a synthetic feasible community, a local community assembled by immigration from a regional pool, and a community generated by repeated diversification. In all three, communities approach marginal stability without developing the near-perfect pairwise correlations predicted by the single-mode picture. Thus, proximity to ecological criticality need not be visible in equal-time pairwise correlations.", "5": "2026-08-22T19:14:05.509487"}
{"0": 461, "1": "arxiv", "2": "https://arxiv.org/abs/2608.20121v1", "3": "Worst-Case Probability Bounds for Finite-Horizon Safety under Moment Uncertainty", "4": "Worst-Case Probability Bounds for Finite-Horizon Safety under Moment Uncertainty. Authors: Renato Loureiro, Torbj\u00f8rn Cunis. Categories: math.OC. This paper addresses the problem of estimating upper bounds on the probability that a dynamical system will enter an undesirable region at some point within a finite time horizon. The primary source of uncertainty lies in the system's initial state, for which only a finite set of moments is known or within a prescribed interval. To tackle this problem, we formulate a measure-based program and propose its relaxation using the moment-sum-of-squares (moment-SOS) framework. The corresponding dual problem is introduced as a functional program, which is subsequently strengthened into a sum-of-squares (SOS) program. Notably, this dual formulation bears a structural resemblance to classical barrier function techniques for certifying system safety, with the key distinction that it yields a probabilistic certificate. The effectiveness of the proposed approach is demonstrated through multiple case studies, including a case involving an object in orbit.", "5": "2026-08-22T19:14:09.678325"}
{"0": 469, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19964v1", "3": "G-MARK: Grounded Multi-Agent Reasoning for Cooperative Driving via Knowledge Graphs", "4": "G-MARK: Grounded Multi-Agent Reasoning for Cooperative Driving via Knowledge Graphs. Authors: Bhavya Gupta, Onat Gungor, Tajana Rosing. Categories: cs.LG. Autonomous driving systems must operate under partial observability, where safety-critical objects may be occluded or visible only to neighboring connected vehicles. Vehicle-to-vehicle cooperation can reduce this uncertainty, but existing cooperative driving methods often compress multi-agent evidence into latent features or hidden multimodal states. As a result, they obscure which agent observed each object, whether the object is visible to the ego vehicle, and how conflicting evidence affects downstream decisions. We propose G-MARK, a grounded multi-agent reasoning framework that converts cooperative object-centric observations into explicit provenance-aware knowledge graphs (KGs). The resulting KGs preserve object hypotheses together with their source attribution, ego-versus-partner visibility, uncertainty, conflicts, spatial relations, and planning-relevant context. G-MARK then derives a shared feature representation from these KGs, enabling lightweight task heads to support object reasoning, motion prediction, control selection, and trajectory forecasting. Compared with the state-of-the-art baseline, GMARK improves occlusion reasoning accuracy by 42.2%, reduces control-selection error by 13.1%, and achieves comparable trajectory-planning accuracy with a 25.6x smaller structured communication payload. Our code is available at https://github.com/bhavyagupta98/g-mark.", "5": "2026-08-22T19:14:09.723882"}
{"0": 470, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19950v1", "3": "Designing Human-mediated AI Guidance: Ready Together for Personalized Family Emergency Preparedness", "4": "Designing Human-mediated AI Guidance: Ready Together for Personalized Family Emergency Preparedness. Authors: Nini Kurashvili, Yana Ivanchenko, Greta Schiavo, Cansu Koyuturk, Dimitri Ognibene. Categories: cs.HC cs.AI cs.CY. Artificial intelligence (AI) systems are increasingly used across domains to provide personalized information, recommendations, and decision support. However, in some contexts, AI-generated information may not be suitable for direct delivery to the final recipient. Instead, it may need to be interpreted, adapted, and communicated by a human who understands the recipient's needs, emotional state, and situational context. Human-AI interaction research has given less attention to situations in which a more knowledgeable human acts as an intermediary between an AI system and a less experienced or less informed recipient. We introduce the human-mediated AI guidance framework and explore it through Ready Together, an AI-supported family emergency preparedness system in which parents mediate AI-generated content for their children. The system is designed to provide personalized guidance and support parents in making emergency preparedness more interactive and understandable through guided activities and family-centered learning. The system design was informed by a qualitative, design-oriented research process involving semi-structured interviews and co-design activities. Findings identified challenges in family emergency preparedness, including difficulty discussing emergencies with children, uncertainty about providing appropriate explanations, and a preference for interactive learning activities. These findings informed the design of an interactive prototype, subsequently evaluated through a pilot study and a heuristic evaluation. Participants responded positively to the personalized recommendations and practical activities. Preliminary findings suggest that human-mediated AI guidance may support context-sensitive family preparedness while preserving parents' responsibility for interpreting, adapting, and communicating AI-generated information.", "5": "2026-08-22T19:14:09.728090"}
{"0": 471, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19902v1", "3": "Bringing analytic rigor to agentic AI for science: The Brain Researcher platform for neuroimaging data analysis", "4": "Bringing analytic rigor to agentic AI for science: The Brain Researcher platform for neuroimaging data analysis. Authors: Zijiao Chen, Nicholas Lu, Xinhui Li, Jocelyn A. Ricard, Ce Ju. Categories: cs.AI q-bio.NC. AI agents can execute scientific analyses, but an analytic output becomes a defensible claim only after alternatives are weighed and the claim is limited to what the evidence supports. Agents may reproduce failures including selective analysis, premature declarations of success and optimization of imperfect criteria. We present Brain Researcher, an agentic research harness operating in a neuroimaging researcher's computational environment under rules for admissible analyses, required checks and claim scope. In benchmarks, Brain Researcher increased first-choice tool-selection accuracy across seven models by 70.2 percentage points (23.3% without it versus 93.6% with it) and verifiable grounding from 4.6% to 22.0%. In collaborator-led and self-evolving studies, multiverse analyses exposed analytic-choice sensitivity, and scientific review classified claims as accepted, qualified, revised, blocked, rejected or deferred. By linking decisions to evidence and provenance, Brain Researcher embeds methodological judgment within the workflow, not after it.", "5": "2026-08-22T19:14:09.734622"}
{"0": 472, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19890v1", "3": "Reliable Neural Collapse Approximation for Open-World Test-Time Adaptation", "4": "Reliable Neural Collapse Approximation for Open-World Test-Time Adaptation. Authors: Jia-Qi Lin, Yuangang Pan, Chang-Dong Wang, Haizhang Zhang, Ivor W. Tsang. Categories: cs.LG. Test-Time Adaptation (TTA) methods aim to bridge the domain gap between the source and target domains. However, traditional TTA methods become ineffective when the label distribution shift occurs, a challenge commonly referred to as an open-world scenario. In this paper, we introduce a new method named Reliable Neural Collapse approximation (ReNC) for Open-World Test-Time Adaptation (OWTTA). Specifically, we leverage neural collapse as a structural prior for reliable target-domain adaptation. Guided by this prior, we justify that the pre-trained classifier weights can serve as the prototypes of the source domain. By measuring the similarity between samples and prototypes, we filter out the Out-Of-Distribution~(OOD) samples for reliable updates. Furthermore, we propose a neural collapse approximation mechanism to refine these prototypes, ensuring they can gradually adapt to the target domain while maintaining the neural collapse structure. Extensive experiments on several open-world benchmarks demonstrate the superiority of the proposed method. Our empirical analysis suggests that ReNC better preserves NC-related properties in the target domain, providing useful evidence for explaining reliable OWTTA and offering new insights for model design. Code is available at https://github.com/JiaqiLin-AI/ReNC.", "5": "2026-08-22T19:14:09.739640"}
{"0": 474, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19866v1", "3": "A 360-Degree Vision Dataset for Learning Yaw Control on GPS-Denied Micro-UAVs in Disaster-Response-Relevant Environments", "4": "A 360-Degree Vision Dataset for Learning Yaw Control on GPS-Denied Micro-UAVs in Disaster-Response-Relevant Environments. Authors: Niklas Voigt, Hartmut Surmann. Categories: cs.CV. This paper presents a novel data-driven approach to camera-based autonomy for micro-drones in GPS-denied, radio-challenging indoor environments. The target application is disaster and emergency response, where micro-UAVs can provide rapid situational awareness in hazardous settings such as firefighting and chemical, biological, radiological, and nuclear (CBRN) incidents while reducing risk for human responders. When the communication link is lost, the micro-drone uses a learned yaw controller to autonomously navigate toward open space, preserving onboard sensor data that would otherwise be lost with the vehicle. A custom micro-drone equipped with a 360-degree camera was used to record diverse industrial, underground, and training scenarios representative of communication-denied field operations. We introduce a preprocessing pipeline that converts equirectangular 360-degree footage into planar front views and dynamically generates image-label pairs for AI training. We then train and compare multiple convolutional neural network variants that predict a continuous yaw command from a single monocular view. Evaluation on a held-out test set confirms the feasibility of the learned yaw-prediction approach. A semi-autonomous real-world test further demonstrates the practicality of the method while revealing key failure modes, particularly reflections and glare.", "5": "2026-08-22T19:14:09.750620"}
{"0": 476, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19838v1", "3": "Specification-delta-driven data governance: an empirical study of the \u00abspec-delta\u00bb as the unit of change in lakehouse data platforms", "4": "Specification-delta-driven data governance: an empirical study of the \u00abspec-delta\u00bb as the unit of change in lakehouse data platforms. Authors: Pablo Ramirez Amador. Categories: cs.AI cs.PL. Spec Driven Development SDD has consolidated the idea that the specification rather than the code should be the primary artefact governing AI assisted work. Tools such as GitHub Spec Kit, and proposals such as Constitutional SDD, have formalised this principle in the software domain, while the executable data-contracts literature has extended it to schema and quality enforcement at run time. Nevertheless, the treatment of the specification delta OpenSpec's core idea that every change should produce a reviewable increment of requirements as the unit of change in data platforms remains empirically unexplored, even though many data-platform changes are contractual (new datasets, service-level agreements, metric semantics, access policies) rather than purely code changes. This work formalises the spec-delta concept, proposes a taxonomy of data platform changes according to their suitability for incremental specification, and defines a controlled experiment comparing a spec-delta-driven workflow against a conventional code pull-request workflow without a delta. The response variables are discovery to deployment time, the density of defects reaching the Silver and Gold lakehouse layers, cross-tool metric divergence, and reviewer cognitive load measured with NASA TLX. The paper explicitly reserves a demonstration-and-laboratory section for instantiation on a real lakehouse environment. The contribution is not a tool but reproducible evidence and an applicability guide that helps to avoid the up front over specification antipattern.", "5": "2026-08-22T19:14:09.760877"}
{"0": 477, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19836v1", "3": "Adaptive Probabilistic Shielding by Learning MDPs for Safe Reinforcement Learning", "4": "Adaptive Probabilistic Shielding by Learning MDPs for Safe Reinforcement Learning. Authors: Astrid Horn Brorholt, Maris F. L. Galesloot, Nils Jansen, Kim Guldstrand Larsen, Christian Schilling. Categories: cs.LG cs.AI cs.LO. Probabilistic shielding is a technique for safe reinforcement learning (RL). Typically, a static observer -- called the shield -- constrains the learning agent's actions to those for which acting safely remains feasible. Traditionally, the shield is computed from the transition probabilities of the underlying Markov decision process (MDP). Thus, this technique is not applicable when the MDP model is not given a priori, which, unfortunately, is the case in typical RL applications. In this paper, we study the problem of computing a shield in the setting where the transition graph of the MDP is known, but the transition probabilities are unknown. Our approach integrates probabilistic shielding with online model learning: as the RL agent explores the environment, we estimate the transition probabilities. From this estimate, we compute a shield. While the shield may be conservative initially, it adapts as the model estimate becomes more precise. Thus, the shield improves in tandem with the RL agent. This paradigm of adaptive probabilistic shielding raises a number of challenges, such as when to recompute the shield and how to balance between exploration and safety during learning. We empirically evaluate multiple variants of this paradigm across several environments.", "5": "2026-08-22T19:14:09.768810"}
{"0": 479, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19819v1", "3": "Large reasoning models for abnormal situation management in safety-critical industrial processes", "4": "Large reasoning models for abnormal situation management in safety-critical industrial processes. Authors: Khalid Alhazmi. Categories: eess.SY. Automation operates safety-critical processes inside their design envelope and leaves abnormal situations to human operators. Mismanagement of these situations is a leading contributor to process-safety incidents and a hindrance to achieving autonomy. Here we show that a general-purpose large reasoning model, with no task-specific training and only the information available to an operator, manages abnormal situations at run time through a bounded, programmatically verified action interface. Across 39 abnormal situations and operating-point changes on a plant-wide industrial benchmark process, the reasoning model maintained the plant within all hard constraints in all 39, while basic regulatory control failed in 15. It matched the plant's expert-engineered advanced control and diagnosed the root-cause fault in 15 of 15 safety-critical situations. Three independently developed models spanning a thirty-fold cost range exceeded the baseline. In a fully auditable evaluation, these results demonstrate run-time abnormal situation management without a human in the loop.", "5": "2026-08-22T19:14:09.777706"}
{"0": 480, "1": "arxiv", "2": "https://arxiv.org/abs/2608.19816v1", "3": "Understanding as an Explicit and Assessable Component of Frontier AI Safety Decisions", "4": "Understanding as an Explicit and Assessable Component of Frontier AI Safety Decisions. Authors: Stephen Barrett, Robin Bloomfield, Alexandra Chiril\u0103, Mamoon Masud, David Meredith Hardy. Categories: cs.CY. Decision makers need sufficient understanding to make good decisions about complex AI systems. However, AI deployment decisions are increasingly made under time-pressure, and this combined with the use of AI generated artefact creation, can mean that the existence of safety cases and system cards may no longer demonstrate that sufficient understanding exists. Our provisional methodology for making understanding explicit and assessable requires the production of an explicit description of 4 objects of understanding (decision, decision-frame, safety justification, system-in-context) and a justification for the adequacy of this understanding. In addition, the methodology provides a mechanism for describing and evaluating the adequacy of the decision-maker representation of this understanding. It builds on recent developments in safety cases using the Assurance 2.0 framework to operationalise the philosophical basis of understanding from Elgin and Arendt. To assess the methodology we trialled two different scenarios. One scenario, which we investigated through role-based analysis, concerned the risk of scheming in the deployment of an AI coding agent in a robotics company and the other scenario was for the higher uncertainty, more decision-critical argument of 'If Anyone Builds It, Everyone Dies' (Yudkowsky and Soares). The trial's central finding, for these two scenarios, is that the methodology could be applied and was found to be generative: we found the analyses that justify sufficiency of understanding (internal coherence, tethering, felicitous falsehoods, external coherence) drives the engineering.", "5": "2026-08-22T19:14:09.786891"}