Physical AI Brief
Daily cross-source signals for the Physical AI supply chain — silicon photonics, CPO, VLA models, humanoid hardware, embodied AI. Three streams, one page, zero filler.
595 items today · 532 arxiv · 1 SEC 8-K · 62 humanoid · 0 CN photonics
01 ARXIV · PHYSICAL AI PAPERS
532 items- arxiv:2610.40655 · cs.CLElucidating the Design Space of LM Data AttributionSteven Yu, Sheng-Yu Wang, Kimia Hamidieh, Antonio Torralba +1
Data attribution methods for Language Models (LMs) are typically evaluated using downstream tasks like fact retrieval. However, existing evaluations vary multiple dataset properties simultaneously, making it impossible to isolate exactly why an attribution algorithm fails. To diagnose these failures, we explore the design space of LLM data attribution, with a primary focus on evaluation. We introduce the Controlled Attribution Benchmark (CATT) for fact tracing, in which we fine-tune LLMs on entirely fictitious knowledge, providing unambiguous ground truth for attribution. This synthetic setup allows us to systematically alter one dataset configuration at a time, such as document redundancy and lexical distractors. With this configurable evaluation, we demonstrate that redundancy and distractors consistently degrade attribution performance. Rather than proposing a new estimator, we examine two design choices that are underexplored: the loss used to compute gradients and how per-token at
benchmark - arxiv:2610.40646 · cs.ROSwapBot: Tool-Centric Robot Manipulation via Interchangeable End-Effectors and Tactile SensingYunchu Zhang, Ray Song, Junhan Jiao, Liyiming Ke +5
Robot manipulation in everyday environments inherently involves substantial tool usage. Seemingly simple household tasks, such as cooking pan-fried tofu, require forceful cutting, delicate brushing, and precise placement using different tools under contact-rich interaction. Thus, the ability to autonomously swap and use task-appropriate tools is important for tool-centric manipulation. We introduce SwapBot, a modular, tool-centric end effector system featuring low-cost adapters that quickly convert common household tools into robot end effectors. Rather than pursuing a single universal gripper, SwapBot adopts a unified design principle that combines a fixed robot-side coupling interface with low-cost, tool-side adapters, enabling compatibility with diverse tools, including articulated ones, and multiple robot platforms. Each adapter integrates compact tactile sensors, supports fast autonomous swapping, and interfaces with an intuitive teleoperation system for efficient demonstration co
manipulationteleoperationtactilegripper - arxiv:2610.40643 · cs.CVLuggage re-identification: A study of strategies towards a real-world applicationQuentin Dhôte, Romaric Audigier
Luggage re-identification (ReID) is a critical task in video surveillance for forensics, security or border management (suspicious or abandoned luggage recognition), as well as luggage management (lost luggage recovery). It enables the association of observations of the same luggage across disjoint camera views in complex environments such as airports, train stations and border crossing. While existing approaches primarily focus on appearance-based cropped image matching under relatively clean conditions, real-world scenarios are characterized by frequent occlusions, cluttered background and crowded scenes which significantly degrade performance. Yet, strategies leveraging the complementarity of multiple shots and views from different video streams is overlooked. In this paper, we study strategies for an end-to-end pipeline that operates from multiple video streams to detect and re-identify luggage. We evaluate different variants on a benchmark dataset acquired in real-world environmen
benchmark - arxiv:2610.40641 · cs.AIFrom extreme weather to agricultural consequences with agentic AIMinghan Yu, Mengyu Dai, Yixin Wen, Haizhao Yang
Extreme weather is forecast days ahead, yet its consequences for agriculture remain hard to predict: a forecast says where a storm will strike, not which crops will suffer or how much is at risk, and the records that could tell span many sources. We present WxCrop, the first agent that judges agricultural hazards from physically interpretable weather indices examined by causal analysis, and the first that combines weather and agricultural information to predict crop loss reliably for past and forecast dates. We demonstrate it in Florida on hurricanes, freezes and dry heat, with a loss model calibrated on insurance records for each, and the same procedure extends to other hazards and areas. Three checks show reliability: the hurricane index detects recorded cyclone days with an AUC of 0.98, recorded losses in held-out years are reproduced to within 2%, and independent production records confirm the county rankings (rank correlation 0.63). Users ask the agent for each crop group's loss a
agentagentic - arxiv:2610.40637 · cs.LGworld2vec: Mechanistic world models through latent dynamics learningTobias Schmidt, Rodrigo Gonzalez Laiz, Krisztián Szuppinger, Steffen Schneider
World models can achieve strong control performance without faithfully capturing the dynamics of their environment, limiting their potential for mechanistic interpretability and scientific discovery. Here, we introduce world2vec to learn mechanistic world models without reward information directly from observations and actions. World2vec learns latent dynamics that faithfully reproduce the true system over long prediction horizons, with at least one order of magnitude fewer parameters than DreamerV3, TD-MPC2 and LeWM. This enables state of the art performance in controlling continuous environments from pixels and accurate system identification across a benchmark of 71 attractors, demonstrating how world models are amenable to scientific discovery.
world modeldreamerv3latent dynamicsbenchmark - arxiv:2610.40632 · cs.ROSymAnt: Grounding Long-Horizon Action Anticipation with Adaptable Symbolic PriorsFnu Aryan, Yaqi Xie, Joseph Campbell, Woojun Kim +5
As robots and intelligent systems increasingly interact with humans, they must anticipate long sequences of future actions from only brief observations. While purely neural models typically require extensive observations to predict future actions and remain difficult to interpret and adjust, symbolic priors offer a structured alternative. To this end, we introduce SymAnt, a novel neuro-symbolic approach that unifies the expressiveness of neural networks with the structured reasoning of scene and knowledge graphs. Given a short video, SymAnt performs a joint-search over both graphs to construct a symbolic context, enabling unified reasoning over spatial relationships, object attributes, affordances, and associated actions. This symbolic context serves as a prior for an iterative refinement process that progressively generates and improves the sequence of future actions. By grounding predictions in symbolic context, SymAnt anticipates long-horizon actions from short observation windows i
knowledge graphiterative refinement - arxiv:2610.40627 · cs.CLShallowDefend: Harnessing Shallow Safety Alignment for Multi-Turn DefenderQiyao Ma, Xiaodong Liu, Weiwei Yang, Mingqian Feng +8
Safety alignment has enabled large language models to reliably refuse harmful requests directly, yet the same request distributed across a multi-turn conversation still succeeds far too often. This brittleness has a mechanistic root: Alignment is shallow, with safety behavior concentrated in a model's first few output tokens, and existing defenses inherit the weakness, supervising the surface form of the response without ever giving the model an explicit safety decision to learn. We take the opposite view of shallowness: if a short span determines the trajectory of an entire response, that span is the highest-leverage place to intervene, and the defender can occupy it first. We propose Assess $\to$ Decide $\to$ Respond, a framework in which the defender assesses the user's intent, commits to an explicit, parseable safety decision, and only then generates a reply conditioned on it, with the commitment trained by reinforcement learning at the payload turn. Because the decision is a token
benchmarkllm-as-judge - arxiv:2610.40619 · cs.CLTestimony from Timelines: Evidence-Grounded Significance Testing of Temporal Clinical HypothesesChengyang He, Wenlong Zhang, Koduvayur Subbalakshmi, Ping Wang
Ask a language model whether sales tend to dip after a product recall, or whether one side effect tends to follow another after a vaccine, and it answers in a fluent sentence that reads like a finding, with no test behind it. Assertions about how events unfold across many cases fill the Web, and each is a temporal hypothesis that only a significance test over the right records can settle. The verifier must specify the null, select the test and decide which records bear on the hypothesis, and any of these choices can go wrong while the conclusion still comes out right. Yet benchmarks for language-model data analysis grade only the conclusion, with the relevant data handed over. We study the problem in health, where a wrong verdict is costly, and present STRATA, a benchmark that grades every decision on the way to a verdict: 8,735 hypotheses over the clinical timelines of vaccine adverse-event reports, with gold plans, evidence sets, p-values and verdicts derived under a released codeboo
benchmark - arxiv:2610.40618 · cs.LGAn ODE-based Sampler for Discrete Diffusion ModelsSamson Gourevitch, Yazid Janati, Eric Moulines, Eric P. Xing +2
The marginals of a continuous diffusion model can be reproduced by a deterministic ordinary differential equation (ODE), known as the probability flow ODE, which provides a foundation for accelerated sampling and consistency distillation. In this paper we address the question whether an analogous formulation exists for discrete diffusion models and answer it affirmatively by introducing exponential residual flows (ERFs). Our framework augments the discrete state with a continuous residual variable for each token and position. These residuals evolve deterministically according to piecewise-defined ODEs whose vector fields depend on the current discrete state and the forward and backward rate matrices of the prescribed diffusion model. The discrete state changes whenever a residual reaches zero. We prove that this construction preserves the marginals of the original diffusion. We derive a practical discretization of our sampler and specialize it to masked and uniform diffusion. The deter
benchmark - arxiv:2610.40614 · cs.CVL3G-World: Real-Time Infinite World Generation with Latent 3D Gaussian MemoryJinhyeok Choi, JoungBin Lee, Jongmin Lee, Minkyeong Jeon +4
Interactive world models generate visual content conditioned on user actions, yet maintaining spatial consistency over long-horizon exploration in real time remains challenging. Existing approaches rely on either temporal context or explicit 3D memory, both of which scale poorly because memory grows with the length of the generated stream. Moreover, explicit 3D memories condition latent video generators by re-encoding RGB renderings, which precludes real-time generation. In this paper, we introduce a Latent 3D Gaussian Memory (L3G) that combines the efficiency of feed-forward reconstruction with a compact, rapidly renderable scene representation. Building on this memory, we propose L3G-World, in which a memory updater and a video diffusion transformer (DiT) alternate autoregressively to generate spatially consistent worlds. The memory updater spawns Gaussians in newly observed regions and refines existing ones, so that the memory grows with scene extent rather than stream length. Each
world modelmemory - arxiv:2610.40611 · cs.AIWhen to Assist and What to Suggest: Joint Gating and Memory-Grounded Hint Generation in One Compact DecoderA. J. A. Jubayer Talukder, Mohammad Ariful Haque
Proactive assistants offer short, memory-grounded hints during a conversation, deciding when to assist and what to suggest. Current systems split this between a gate and a larger generator; we evaluate both together on 28,058 decision points. Stage-wise generation scores are oracle-gate ceilings: behind a real gate, end-to-end exact match is 0.2619 against a ceiling of 0.2971, and end-to-end precision is 0.2387, leaving substantial headroom. A gate pays off through precision and compute, whose saving falls from 4.39x with an 8B generator to 1.38x with a 1B one. The memory profile drives hint quality: removing it costs 24.7 exact-match points (rescaled BERTScore 0.57 to 0.35), while retrieval and a larger generator add nothing on exact match. One 1B decoder can do both, gating from an intermediate layer and generating in the same pass. At its final checkpoint it matches or exceeds the two-model pipelines on automatic end-to-end measures with 1.31-1.76x less wall-clock time than the matc
memory - arxiv:2610.40607 · cs.CVFight World Model: Customizable Combat through Visual Interaction ProxiesJiahua Li, Zixu John, Tom Zhong, Fuping Wu +4
A playable, programmable game world must maintain persistent state, enforce executable rules, and support extensible gameplay mechanics. Existing video world models can generate plausible combat, yet struggle to keep hit reactions and health updates consistent with game rules. We introduce Fight World Model (FWM), which separates rule-governed combat execution from video rendering. An executable engine maintains state and resolves interactions. A deterministic compiler encodes these outcomes and attempted actions in a spatially and temporally aligned interaction proxy. The proxy aligns action and skill windows, brief hit reactions, and persistent health by fighter identity and timing. These cues can be controlled separately to render interactions under modified rules. Identity-bound references preserve character appearance across side switches. We train FWM on FightWorld Corpus, spanning 52 skill effects, 187 character-appearance variants, and 15 arenas. Without retraining the renderer
world modelpersistent statearena - arxiv:2610.40589 · cs.AICollaborative AI agents for autonomous experimental research with solid-state quantum spinsEl Mustapha Mansouri, Eikichi Kimura, Masato Koga, Keigo Arai
An autonomous research laboratory could conduct an entire study, from formulating questions to preparing manuscripts. Artificial intelligence (AI) offers one route towards this goal. Recent AI systems take on several research roles, but may draw conclusions that the evidence does not support. Here, we present an autonomous research laboratory in which agents based on large language models assess one another's work and can stop or redirect investigations. They can generate and reject research questions, require separate theoretical checks and conduct experiments designed to challenge their claims. We demonstrate these capabilities by investigating ensembles of quantum spins in diamond. The agents derived the minimum number of calibration settings needed to predict responses to a defined family of microwave pulse sequences. They checked the result computationally and designed experiments to distinguish competing predictions. An initial experiment agreed with the theoretical predictions,
ai agent - arxiv:2610.40588 · cs.AIPre-Crime for Agents, at the Last Second: Probing When a Recruited Agent Gives InLeonidas Raghav
When AI agents share channels and credentials, one agent can press another to break its operator's rule, a first step towards collusion. Output filters check only what an agent has already written. Can an overseer tell beforehand that an agent is about to give in? We show that a linear probe on the agent's activations can, in a controlled multi-agent setting. Across six scenarios, a peer presses a target for up to 20 rounds to reveal a secret, such as a deploy token; the target often refuses, then gives in. Read at each model call before its output is written, the probe picks out the one that leaks. On held-out Qwen3-32B episodes, at a threshold set to flag 5% of benign calls, it catches 82% of leaking calls. At a similar flag rate, a zero-shot decision-model monitor catches 26%. On gpt-oss-20b, a probe on its reasoning reaches AUROC 0.91. A bag-of-words classifier trained on the same calls comes close, so most of the gain comes from training on labels, not activations. The probe also
agentai agentmulti-agent - arxiv:2610.40578 · cs.CVVeriCUA: Verification as Agentic Exploration for Computer-Using AgentsQiushi Sun, Da Yu, Alexander Knop, Amer Sinha +4
Building and scaling computer-using agents (CUAs) across desktop, mobile, and web environments requires reliable task-completion verification for data curation, evaluation, and reinforcement learning. Typical automated approaches prompt vision-language models (VLMs) to judge trajectories or generate programs to inspect application state. However, both rely on partial views of complex digital environments. VLM judges are restricted to supplied observations, while generated programs depend on accessible state. These views can omit decisive evidence or leave completion judgments unsupported. Tailoring separate verifiers to different platforms and access settings adds further engineering cost. We present VeriCUA, a platform-agnostic framework for verification through agentic exploration. It treats verification as dynamic evidence gathering and assessment within a shared procedure for online and offline settings across platforms and observation modalities. After inspecting the available con
agentic - arxiv:2610.40576 · cs.CLPIVOT: Progressive Intervention via Online Teacher CorrectionZhihao Yang, Boyuan Zheng, Puhui Tan, Yukun Chen +6
Reinforcement learning from verifiable rewards (RLVR) is limited by sparse outcome-level feedback on hard reasoning tasks: when all rollouts in a group fail, rewards provide no learning signal. Stronger teachers can provide additional guidance, but this raises a central question: how can such signals be incorporated into RLVR to support successful rollouts while remaining learnable for the student? We distinguish two complementary mechanisms: trajectory-level intervention, which improves access to reward-bearing trajectories, and token-level supervision, which provides dense targets to guide student learning. Our controlled experiments show that either mechanism alone can remain insufficient, motivating their combination. We therefore introduce PIVOT, which inserts teacher-generated corrections into student rollouts and applies token-level KL supervision to these segments alongside the RLVR objective. Beyond their combination, we systematically analyze where to intervene and how to sup
benchmark - arxiv:2610.40574 · cs.CLEvery Token Needs Its Own Teacher: Multi-teacher Adaptive On-policy Self-distillationYifu Luo, Chuang Ma, Zeyu Chen, Enkelejda Kasneci +2
On-policy Self-distillation (OPSD) has proven effective for post-training large language models by constructing a teacher conditioned on privileged references. However, existing approaches keep the teacher fixed either throughout training by consistently exposing it to the full reference, or within each response by adaptively varying partial reference exposure only across responses. In this paper, we argue that neither of them is sufficient. Instead, every token needs its own teacher, as the teacher preference can vary token by token within the same response. Through a teacher-intervention pilot study, we validate that the teacher preference differs across tokens, and reveal that a preferred teacher can be identified for each token through majority voting and KL-based selection. Motivated by these findings, we propose Multi-teacher Adaptive On-policy Self-Distillation (MTA-OPSD), the first token-level adaptive OPSD method. Our method constructs a multi-teacher pool consisting of candid
post-training - arxiv:2610.40565 · cs.ROGym-Anything for Robotics: Scaling physical simulations with digital agentsPranjal Aggarwal, Sean Welleck
Realistic, fast robotics simulations could enable large-scale training of embodied AI models that transfer to the real world. However, despite decades of progress, available simulations cover only a narrow range of real-world tasks and physical phenomena. The core bottleneck is that creating better simulations requires significant expert effort and therefore does not scale. To address this, we introduce Gym-Anything for Robotics, a framework in which coding agents create realistic robotics simulations. Our core insight is that simulation creation is a coding agent task, so improvements in coding agents translate directly into richer and more faithful simulations. We select simulations based on U.S. GDP data to cover economically valuable physical work, define what makes a good simulation and a good task, and develop an adversarial creation-audit loop that checks each simulation and task against these definitions. Using this pipeline, we build RoboWorld, a set of 25 simulations with 25
vision-language-actionembodiedagentbenchmark - arxiv:2610.40559 · cs.ROA Tutorial on Shape Servoing of Deformable Objects-Part I: FundamentalsIgnacio Cuiral-Zueco, Alessio Caporali, Miguel Aranda, Gianluca Palli +2
In robotics, shape control (or shape servoing) refers to the automatic manipulation of a deformable object to achieve a desired shape. It involves a feedback loop that exploits the relation between robot motion and changes in the measured shape. This relation, known as the interaction Jacobian or, in this context, the deformation Jacobian, depends on the object's mechanical properties, geometry, and boundary conditions (e.g., where the object is grasped). Because these factors are often difficult to model accurately, the deformation Jacobian can be estimated locally from observed shape changes produced by commanded robot motions, without requiring a fine-tuned mechanical model. This article is the first of a two-part tutorial on data-driven Jacobian-based shape control of deformable objects. It introduces the central elements of the approach: the formulation of the shape-control problem, shape representation, estimation and online update of the deformation Jacobian, and a standard cont
manipulationgrasp - arxiv:2610.40556 · cs.AIAdvancing Clinical Genomic Analysis with a Skill-Centred AgentQiaoyu Zheng, Weike Zhao, Yanjie Fan, Ye Cao +9
Clinical genomic analysis requires experts to integrate evolving guidelines, databases, and computational evidence. As sequencing has become increasingly accessible, this complex and labor-intensive process has become a major bottleneck to the timely use of genomic information in clinical decision-making. To fill this gap, we developed ClinGeX, a large-language-model agent that composes 105 curated executable skills through task-specific retrieval and dependency-aware orchestration. On 223 temporally held-out benchmark cases, ClinGeX achieved 71.7\% macro-averaged task-level concordance with expert reference classifications, compared with 45.7\% for the strongest comparator. In a three-condition, counterbalanced human-AI collaboration study, 45 clinical genomic professionals assessed 135 clinically derived cases under routine (no AI assistance), DeepSeek-assisted and ClinGeX-assisted conditions. ClinGeX assistance increased raw case-level concordance across criterion-level variant anal
agentbenchmark - arxiv:2610.40551 · cs.AIRamanujan Challenge Results: AI-Assisted Evaluation Pipeline for Multi-Format ProofsAlan Sorani, Ashvni Narayanan, Michael Shalyt, Tali Monderer +7
As AI's mathematical capabilities rapidly improve, meaningful benchmarks are shifting from university-level problems and math olympiads to research-level mathematics. The Ramanujan Challenge is a benchmark comprised of research-level problems in number theory, including unpublished results and open conjectures. It presented concrete formulas for mathematical constants whose correctness can be checked numerically, while their proofs may require non-obvious mathematics. Here, we analyze the Challenge submissions, measuring the relative success of different AI models, Computer Algebra Systems (CAS), and formal tools. We find correct submissions that are conceptually distinct from the official solutions, and a proof of one of the published open conjectures. Evaluating the submissions exposed a second challenge: judging complex proofs reliably at scale. We introduce an AI-assisted evaluation pipeline for natural language proofs, CAS code, and Lean formalizations that assesses semantic and l
benchmark - arxiv:2610.40545 · cs.AIRetrieval as Traversal: Value-Indexed Memory for Stateful Long-Context AgentsDaniel Strauß, Zhenshan Bing, Alois Knoll
Long-context models repeatedly solve closely related retrieval problems. Attention or global retrieval can retain large stores, but each forward pass again searches for what matters. Recurrent compression avoids this search only by deciding earlier what information to preserve. Neither mechanism directly exploits that useful memories at neighbouring retrieval steps may substantially overlap. We propose an alternative in which retrieval is a persistent local traversal through an address space. We separate a memory's content from its address. Utility is defined by the information revealed by retrieving the content, while geometry determines only how cheaply that information can be reached. This leads to a cost-regularised indexing objective that preserves information value inside local neighbourhoods while charging for every memory made locally accessible. Its marginal-value form gives both attractive and exclusion pressures. Thus, complementary memories should be co-retrievable when the
memorylong-context - arxiv:2610.40541 · cs.AIOn the Fundamental Limits of Prefill-Based Decoding Length PredictionHaowen Shi, Xuzhe Yan, Fahao Chen, Peng Li
Predicting how many tokens a large language model (LLM) will decode, using only prefill-phase hidden states, is a key primitive for inference scheduling. We show that current benchmarks conflate predictor quality with task hardness: when the target model, dataset, or sampling configuration changes, reported accuracy collapses even though the predictor is unchanged. We trace this instability to an overlooked error source---the inherent stochasticity of autoregressive decoding---and derive bounds that every prefill-side method must obey. Our analysis treats decoding as a stopped stochastic process. The per-step stopping probability of a response is not constant but ramps up over hundreds of steps, so the decode length inherits a central limit theorem from the many local decisions that determine where a response ends; this yields per-prompt normality with an explicit error term, together with an affine variance law $σ=αμ+b$ whose slope is the coefficient of variation of the number of resp
benchmark - arxiv:2610.40539 · cs.ROPhysVLMM: Interaction-Conditioned Physical Belief Maps for Articulated Objects in Robot Scene MapsDibyendu Ghosh
Mobile robots must open doors and drawers, yet their scene maps record what each object is and whether it has moved, not how it will respond to a pull. This paper proposes PhysVLMM, a method that maintains a physical belief for every articulated part of a Vision-Language-Motion Map and answers a planner's queries about it, with two algorithms: a map update and a query. Watching a part move provably cannot identify its dynamics, and a measured pull identifies only the pulled part, up to one direction. PhysVLMM therefore (i) keeps that direction as a posterior instead of a point estimate, (ii) passes the evidence of each pull in closed form to the untouched look-alike parts of the map, and (iii) answers whether an action will succeed at a stated risk, or abstains. On real doors and drawers in 18 homes, it identifies the joint type in 86% of held-out pulls (48% for a learned single-view method) and predicts the response to a pull with 41% lower error than differentiable simulation (CRPS 0
action-conditioned - arxiv:2610.40537 · cs.RODAWN: Diffusion-based Alignment and unfastening Workflow NetworkHemant Kumar, Piyush Goenka, Nikhil Chopra
We present DAWN (Diffusion-based Alignment and unfastening Workflow Network), a hierarchical robotic framework designed to perform real-world, unfastening tasks through multi-modal sensory integration. Autonomous robotic unfastening is crucial for improving recycling processes, enhancing safety, and reducing operational costs; yet, it remains predominantly manual due to inherent challenges in precise screwdriver engagement with fasteners, detailed force execution, and robustness to environmental variability. To address the complexity of learning long-horizon, contact-rich manipulation tasks, we subdivide the unfastening task procedure into two sequential stages: (1) alignment & insertion, and (2) unfastening. DAWN consists of three key components: AlignNet, a diffusion-based policy ensuring precise screwdriver tip engagement; LatchNet, a multi-modal classifier reliably detecting successful stage transitions; and UnfastNet, which achieves stable, force-controlled screw removal. Experime
manipulationfranka - arxiv:2610.40532 · cs.ROBURGER: Benchmarking and Understanding Robotic Generalization via Embodied DrawingXiangyu Xu, Bangshuai Peng, Ziyi Liu, Mengjie Zhao +4
How effectively can vision-language-action (VLA) models learn diverse language-conditioned behaviors as training data scales? Investigating this question requires large and varied robot datasets, whose collection on physical robots remains costly. We introduce BURGER, a language-conditioned robotic drawing benchmark with a scalable, prompt-driven pipeline for generating continuous drawing trajectories, both in simulation and the real world. In BURGER, an agent starts from the same blank canvas and controls a robot arm to draw the object or spatial composition specified by language, while each instruction allows an open-ended variety of valid drawings and stroke orders. Using 28,200 simulated demonstrations spanning 282 drawing tasks, we evaluate five representative VLA models across data scales, language grounding, and multimodal behavior. BURGER supports fine-grained analysis of language grounding, spatial composition, and action generation, revealing distinct failure patterns across
vision-language-actionvlavla modelembodiedfrankaagent - arxiv:2610.40530 · cs.CVPhysCount: Do MLLMs Count Visual Appearances or Physical Entities?Lidong Lu, Jiaxin Lv, Liangrui Dong, Xinyi Mao +4
Images reveal appearances, but counting often asks about physical objects. A single entity may be visible directly, in several reflections, or through glass, so detecting every visible occurrence does not determine how many entities are present. Existing visual counting benchmarks rarely separate these two operations. We introduce PhysCount, a benchmark for evaluating whether multimodal large language models (MLLMs) can distinguish what appears in an image from what physically exists in the scene. PhysCount contains 546 real-world images and 4,595 human-written questions organized into 651 target-specific groups. For each group, annotators locate the relevant appearances, label how each is observed, and identify appearances of the same entity. Four complementary question families test the enumeration of appearances, the enumeration of physical entities, image-formation-specific counting, and appearance--entity correspondence. We measure both accuracy on each family and the proportion o
benchmark - arxiv:2610.40528 · cs.LGNEOPIC: A Neural Operator Framework for Particle-based Kinetic Plasma SimulationsSriramkrishnan Muralikrishnan, Chelsea Maria John, Jeffrey Kelling, Michael Bussmann +2
We introduce NEOPIC, a framework for learning the field operator in particle-based methods applied to kinetic plasma simulations. Its key novelty is a neural operator that acts directly on the empirical measure of the particles, rather than on a grid, and is evaluated at the particle positions. No mesh or tree is needed and, unlike grid-based neural operators, no deposition onto a mesh takes place, so the particle noise that dominates the inputs of particle simulations is not aliased. Since it replaces only the field solve, it serves as a plug-and-play field solver for particle-based methods: it can be trained on data from particle-in-cell, particle-in-Fourier or tree codes, and combined with any time integrator, including implicit schemes. We prove that the discrete evaluation coincides with the continuous operator at any number of particles (representation equivalence) and derive an error analysis that separates sampling, learning and time-stepping errors: up to the learning error, N
benchmark - arxiv:2610.40518 · cs.CLIndicIPR-QA: A Benchmark for Extractive Question Answering on Indian IPR LawsM. Mohaiminul Islam, Muhammad Abulaish
Legal Question Answering (QA) over statutory texts is a challenging task in Natural Language Processing (NLP), especially in low-resource settings with complex legal systems. It requires answers that are precise and grounded in the law. However, to the best of current knowledge, no existing extractive QA benchmark provides exact answer spans from statutory texts across all seven Indian Intellectual Property Rights (IPR) Acts. This paper presents IndicIPR-QA1, an extractive QA benchmark for Indian IPR laws. The proposed dataset contains 4,462 QA pairs from 1,116 statutory contexts across seven Acts. Questions and answers are generated using GPT-4o and validated through nine automatic quality checks, single and multi-model LLM evaluations, and a 200-pair human audit. Section-level partitioning prevents source-context overlap across the training, validation, and test sets. Further analysis examines answer span length, lexical overlap, and multi-clause structure across the corpus and indiv
benchmark - arxiv:2610.40503 · cs.AIAlgebraic World Model: From Exploration to ProofsGuancheng Pan, Qixin Tan, Yang Yuan, Andrew Chi-Chih Yao
AI has made remarkable progress in mathematics, solving challenging problems with machine-checkable proofs in Lean, yet typically relying on predefined mathematical definitions and formal specifications. This raises a natural question: can it reason rigorously about an unknown world? We introduce the Algebraic World Model (AWM), a framework that turns interaction with unfamiliar environments into explicit mathematical models supporting formal reasoning. AWM explores the environment and develops executable Python hypotheses, continually testing and revising them against observations. It then structures these hypotheses into typed objects and shared rules, simplifies their structure through category-theory-guided transformations, and formalizes the resulting model in Lean 4 to prove reachability and impossibility. AWM can formally prove that a goal is unreachable, even beyond explored trajectories. On two impossible game levels, it constructs closed, kernel-checked Lean proofs excluding
world model - arxiv:2610.40495 · cs.CVLucidLyra: Feed-Forward Generative 3D Scene Reconstruction and Joint Understanding via Multi-Task DistillationOliver Hahn, Rodrigo Marcuzzi, Sangeetha Grama Srinivasan, Alessandro Burzio +6
We present LucidLyra, a framework that makes generatively reconstructed 3D scenes directly queryable for semantics and instances in a feed-forward manner from video latents. While generative 3D reconstruction can synthesize virtual environments from a single image, downstream applications in robotics and physical AI require not only visual realism but also access to scene semantics and object identities. Existing 3D understanding approaches, however, require captured multi-view scenes, have limited generalization capabilities, or rely on expensive per-scene processing and optimization. To bridge this gap, our method distills complementary knowledge from a camera-controlled video diffusion model and frozen 2D vision foundation models into a unified model to jointly infer 3D Gaussians encoding geometry, appearance, open-vocabulary semantics, and class-agnostic instance cues. The resulting scenes support novel-view rendering, open-vocabulary segmentation, and object grouping without per-s
benchmark - arxiv:2610.40492 · cs.AIAgentic University: Collaborative Agents for Personalized LearningJingru Fan, Tianle Zhou, Yunjie Zhou, Zihao He +9
Large language models support personalized tutoring, content generation, and adaptive learning paths. However, capability development requires these resources and interactions to form a coherent learning process that evolves with the learner. A central challenge is to align goals, activities, and assessment while using learning evidence to decide what should follow. We introduce Human-Participatory Multi-Agent Organization and instantiate it in Agentic University (AU). In AU, learners participate as organization members and work with specialized agents to maintain a revisable Capability Program. The program connects personal goals, learning activities across modalities, and layered assessment, while learner feedback and performance guide subsequent support. Our evaluation shows advantages in program organization, assessment design, and revision quality. Users achieved larger mean assessment gains (12.19 versus 5.61 points on a 50-point scale) and reported stronger cognitive support. Th
multi-agentagentic - arxiv:2610.40488 · cs.CLrTaste: Do Research Agents Have Taste? Evaluating Early Stage Research JudgmentArt Kanke
Do research agents have taste, the judgment of which experiments are worth running before their outcomes are known? We give six frontier models seven recent NLP papers. Each model, working as a research agent alone or as a team, proposes minimum viable experiments (MVEs), forecasts whether each will pass a threshold it sets and ranks them. A separate executor then attempts every MVE, including those ranked last, and three model auditors judge the results. Of 415 MVEs, 194 are judged sound, valuable and new, but fewer than half of these (73) are also judged beyond the seed paper and non obvious. The agents' first choices match the auditors' most valued experiment more often than a random pick, but are not shown to improve on choosing the experiment written first. Their forecasts beat a base rate but are not measurably more accurate than those of a fresh instance of the same model given only the paper and registrations. Teams cost more without a clear overall improvement in proposal qual
agent - arxiv:2610.40487 · cs.CVMetaRefine: Score Refinement with Multi-Level Features for Image AssessmentChuanzhi Xu, Chengkun Yue, Haolan Guo, Xuanhua Yin +4
An image's original dimensions, aspect ratio, and composition provide important cues that can affect its technical and aesthetic quality. Image assessment methods use resizing, cropping, padding, or patch sampling to accommodate different input sizes. However, these operations can alter or obscure original geometric and compositional cues. Existing assessment models also encode rich multi-level visual evidence that can be further extracted and exploited to improve existing models. This motivates us to combine visual evidence with original-image geometry for score refinement without changing the original scoring pipelines. In this paper, we introduce MetaRefine, a metadata-aware score refinement architecture whose lightweight residual heads predict score corrections from geometry, multi-level visual features, and the frozen model's original score. Across 100 configurations on 21 datasets, MetaRefine improves PLCC in 88 configurations and SRCC in 85, with average relative gains of 7.55%
benchmark - arxiv:2610.40483 · cs.LGPreserving Utility in Continual LLM Unlearning via Orthogonal ProjectionsJuan Belieni, Ana Carolina Erthal, Eliezer de Souza da Silva, Diego Mesquita
Unlearning concerns the task of removing the influence of a sensitive portion of the training data, such as protected, copyrighted or private text, from a model without retraining it from scratch. While unlearning requests happen continuously during a model's lifetime, the vast majority of LLM unlearning methods are ill-suited for iterated removal, yielding steep decay in forgetting and utility upon sequential unlearning. This decay stems from interference between successive updates, as each new deletion can undo earlier ones or push the model further along directions already degraded by them. To address this, we introduce Continual Orthogonal Projection (COP), a gradient-projection scheme that restricts each update to the orthogonal complement of activation subspaces cached from prior forget sets. This approach provably guarantees first-order non-interference with earlier deletions. COP stores only a low-rank basis per layer, requiring neither past gradients nor previous forget sets,
benchmark - arxiv:2610.40482 · cs.AIFIRME: Registry-Anchored Memory Graphs for Evolving, Heterogeneous Investor KnowledgeYunbai Zhang, Yunyi Gao, Han-Tai Shiao
Financial AI workflows that reuse investor knowledge need persistent memory that integrates observations across sources and periods without losing identity or provenance. In a knowledge graph, recurring observations may use different surface forms, fragmenting relationship histories and preventing evidence from accumulating. We introduce Fusion and Identity Registry for Memory Evolution (FIRME), a registry-anchored memory graph that reconciles incoming observations to persistent canonical relationships and retains their source, time, and context. Within periods, rule-based fusion links investor-call discussions to portfolio cash-flow signals, directly or through a fixed product-topic ontology. Across periods, canonical identities and persistent aliases direct recurring observations to the same relationship history; deterministic updates summarize reinforcement, inactivity, and reactivation. We evaluate FIRME on proprietary investor-call and cash-flow records and on S&P 500 earnings-cal
memorypersistent memoryknowledge graph - arxiv:2610.40480 · cs.CVAVPS-360: A Spatial Audio-Visual Benchmark for Instance Segmentation and Tracking of Sound-Producing PersonsYifu Sun, Parthasaarathy Sudarsanam, Tuomas Virtanen, Archontis Politis
Audio-visual instance segmentation (AVIS) segments and tracks individual sound-producing objects in videos. Existing benchmarks predominantly use monaural audio, which lacks inter-channel cues for spatially distinguishing concurrent sound sources. We introduce AVPS-360 (Audio-Visual Person Segmentation 360), a spatial audio-visual benchmark bridging AVIS and sound event localisation and detection (SELD). Built upon STARSS23, it pairs first-order Ambisonics audio with synchronised 360° video across 230 recordings totalling 10.5 hours. Recordings average 164.0 seconds. The dataset captures intermittent sound production in real-world interactions dominated by speech, alongside laughter, clapping, and footsteps. It provides 31,678 manually reviewed instance masks linked to source identities and sound-event labels. We also present a baseline that integrates spatial audio with visual features for frame-level segmentation and identity association across frames. Experiments on AVPS-360 show th
benchmark - arxiv:2610.40476 · cs.AITraining language models in simulation for psychiatric intakeSamuel Schmidgall, Tim Strother, Alex Bijamov, Min Woo Sun +52
In psychiatric practice, the clinical intake interview remains a primary diagnostic instrument, requiring the clinician to establish rapport while disentangling complex, overlapping presentations. Here, we show that multi-turn reinforcement learning (RL) in simulation environments improves language model performance on structured text-based psychiatric intake, spanning history-taking, differential diagnosis, and initial management planning. In a randomized, double-blind crossover study ($N = 60$ matched scenario pairs, 120 encounters) comparing the RL-trained agent against its identically prompted base model, 32 board-certified psychiatrists role-played standardized patients while a separate panel of 15 independent, board-certified psychiatrists scored the de-identified transcripts. The trained agent outperformed the base model in 76.7% of matched scenario pairs ($p < 0.0001$), with gains on an adapted clinical-examination rubric ($d_z = 0.77$, $p < 0.0001$), patient-centered communica
agent - arxiv:2610.40467 · cs.CVInfantNet: A Large Scale Dataset for Infant Body Pose and Shape EstimationZhipeng Hui, Jiahao Wang, Xun Dong, Guofeng Zhang +12
Infant pose and shape estimation is essential for applications in childcare, developmental monitoring, and medical diagnosis. However, existing methods and datasets are largely designed for adults, and direct transfer to infants fails due to substantial differences in body proportions, articulation limits, and frequent self-occlusion. To address this gap, we introduce InfantNet, the largest real-image infant dataset to date, comprising 108,902 RGB images of infants aged 6-18 months. Each image is annotated with 2D keypoints, and a curated subset of 11,642 images additionally includes 3D pose and shape annotations with full SMIL parameters, verified through multi-view rendering and reprojection-error filtering. We use an iterative annotation pipeline to ensure high fidelity across both 2D and 3D labels. InfantNet establishes a large-scale, comprehensive benchmark for infant 2D keypoint detection and 3D pose-and-shape recovery. Baseline experiments demonstrate that state-of-the-art adult
benchmark - arxiv:2610.40455 · cs.CVA Unified Benchmark for Aerial-Enhanced Autonomous Driving: The KITScenes Aerial DatasetDaniel Lengerer, Richard Schwarzkopf, Jonas Merkert, Fabian Immel +7
Aerial imagery provides complementary spatial context for autonomous driving by exposing road structure beyond the immediate perception range of onboard sensors. Yet map construction from aerial imagery, aerial-assisted online HD mapping, and cross-view localization are typically studied on separate data and benchmarks. We introduce KITScenes Aerial, an extension of the KITScenes Multimodal dataset that augments synchronized onboard sensor data, vehicle poses, and rich Lanelet2 map annotations with spatially associated aerial imagery and higher-resolution neural representations, providing a common basis for a variety of research tasks. Our mapping evaluation focuses on all 53 KITScenes map classes across different model designs, spatial ranges, and aerial resolutions, while retaining compatibility with the widely used three-class setting. Representative baselines show that aerial context is particularly beneficial for fine-grained and extended-range mapping, more than doubling 53-class
benchmark - arxiv:2610.40454 · cs.CLGlance: The Generator Is Its Own RerankerAdilet Metinov, Chingiz Arziev
Retrieval-augmented generation commonly serves separate models for reranking and answering, duplicating model residency and passage processing. We introduce Glance, which equips the generator with a lightweight decision adapter and pointer head to score candidate passages from a shared prefill. Short relevance probes branch from the cached context, producing one score per passage without separately prefilling each query-passage pair. A single resident 9B backbone supports generation, general decisions and reranking, with approximately 0.1 GiB of additional memory per decision skill. After 39 minutes of MS MARCO fine-tuning, Glance achieves nDCG@10 of 0.741 on SciFact and 0.345 on NFCorpus, compared with 0.721 and 0.326 for bge-reranker-v2-m3. On 300 HotpotQA questions, its selected context improves answer F1 by 0.045 over cross-encoder-selected context under the same base generator (paired bootstrap 95% CI [0.014, 0.076]). In a scaling study, reranking quality keeps rising up to 200 ca
memoryretrieval-augmented - arxiv:2610.40449 · cs.LGReinforcement learning for control of partial differential equationsNicolò Botteghi, Matteo Tomasetto, Andrea Manzoni
Controlling dynamical systems governed by partial differential equations (PDEs) remains a challenging task in scientific computing, owing to high-dimensional state spaces, distributed actuation, nonlinear dynamics, and stringent real-time constraints. Reinforcement learning (RL) has emerged as a promising paradigm for synthesizing feedback controllers that adapt to varying operating conditions and can be deployed efficiently in real-time. This review provides a unified perspective on RL and PDE-constrained optimization, and surveys model-free, model-based, and multi-agent RL approaches to PDE control, with particular emphasis on sample efficiency, generalization across parametric variations, and scalability to high-dimensional problems. Two demonstrative examples, the stabilization of a chaotic system and a flow control problem, illustrate the performance and pitfalls of RL agents in practice. Finally, we discuss evaluation and benchmarking practices, current limitations, and open rese
multi-agentbenchmark - arxiv:2610.40444 · cs.AIPrefix Executability: Evaluating Tool-Using Agents Beyond Final SuccessAmir H. Rezaeian, Weiyi Sun, Yassine Benajiba, Dan Roth
Tool-use benchmarks often score agents by final task success, but terminal outcomes can obscure whether the intermediate tool-call trajectory was executable. We propose prefix executability as a trajectory-level evaluation protocol for tool-using agents. The protocol measures whether every prefix of a predicted tool-call sequence remains executable under tool schemas, state preconditions, and data dependencies, and summarizes reliability using Prefix Executability Curves (PEC), First-Error Depth (FED), Trajectory Executability Rate (TER), horizon-normalized AUC-PEC, and Recovery Rate (RR). We instantiate the protocol on stratified samples from BFCL-v3 multi-turn and ToolSandbox, and on the complete Airline/Retail base splits of $τ^3$-bench. We observe that outcome-only and prefix-level rankings diverge: agents whose execution is conditioned on an unverified plan often achieve higher point-estimate final success than non-plan-conditioned baselines while substantially lowering TER and AU
tool-usebenchmarkevaluation protocol - arxiv:2610.40443 · cs.LGFramework-Conditional Open Materials GenerationMaya M. Martirossyan, Hillary Pan, Philipp Höllmer, Stefano Martiniani
Theoretical predictions of novel materials rely on a variety of techniques, from data-mining-based ion substitution methods to fully de novo generative models. Each approach imposes specific constraints and encounters unique challenges when extrapolating beyond the distribution of known materials to discover new ones. The foremost goal of materials generation is to produce stable, unique, and novel (S.U.N.) structures. Predicted structures that meet these criteria, however, typically result from novel chemical compositions mapped onto known structural prototypes or frameworks. We leverage this heuristic insight to present a new generative paradigm for the de novo prediction of inorganic crystalline materials constrained by their structural framework---i.e., the unit cell shapes and respective particle positions. Our method, termed OMatG-FC (Open Materials Framework-Conditional Generation), learns via a combination of discrete flow matching on the atomic species and stochastic interpola
benchmark - arxiv:2610.40430 · cs.ROImproving Fine-Grained Manipulation through Demonstration Expansion and Execution DesignJingxu Gu, Dongze Lian
Fine-grained manipulation requires a robot policy to follow instructions, remain robust to visual variations, and translate predicted actions into precise robot control. We present Demonstration Expansion and Execution Design (DEED), a framework for adapting a pretrained vision-language-action policy to the five MetaFine tasks through structured demonstration expansion and execution design. DEED generates demonstrations for different instructions from the same initial scene and renders each physical trajectory under different camera positions and lighting conditions while preserving its original states and action targets. A hierarchical sampling strategy balances instructions, trajectories, and visual conditions during supervised adaptation. At deployment, DEED converts predicted action chunks into robot commands using inverse kinematics, gripper mappings selected for each task, and closed-loop replanning. DEED achieves an official Overall score of 85.55%. Fixed-policy evaluations show
vision-language-actionmanipulationrobot policygripperpolicy evaluation - arxiv:2610.40427 · cs.ROsocx: GPU-Accelerated Chance-Constrained Stochastic Optimal Control using JAXTanmay Dokania, Yashwanth Kumar Nakka
In this article, we present socx, a GPU-accelerated solver for lifted convex reformulations of linear stochastic optimal control under chance constraints, implemented in JAX. We demonstrate that by exploiting the sparse and independent structure of the problem, distributed optimization can reduce it to simpler problems: independent linear-quadratic tracking (LQT) problems and Euclidean projections onto sets. These simpler problems can be solved in parallel and then coupled using the constraints. To validate our solver, we run it on a hover-linearized 3D quadrotor and show a speed-up of up to $98\times$ over MOSEK and up to $2224\times$ over SCS. socx also solves horizons up to $N=1120$, whereas MOSEK runs out of memory at $N=320$. We present ablations to isolate the key components that enable the speed-up.
memory - arxiv:2610.40415 · cs.LGOCCAMS: A Multi-Task Learning-Based Investigation of the CMB Power SpectraAnkit Mandal, Rahul Shah, Md. Wali Hossain, Somasri Sen +2
How many latent variables are needed to fit the cosmic microwave background (CMB) temperature and polarisation power spectra when a neural network is given no cosmological model, and to what extent do they map onto the standard cosmological parameters? We use multi-task learning with variational autoencoders trained on CLASS-generated spectra (TT, TE, EE) over $2\le\ell\le2500$. No cosmological model is built into the network, and the number of latent variables, $L$, is varied over $L=5$, $6$ and $7$. Their fits are compared to the Planck 2018 data using $\hatχ^2$, AIC, BIC and Bayesian evidence. Their correspondence with cosmological parameters is measured with normalised mutual information post-training. The loss weights are set within a single training run by uncertainty weighting, with the weight of the disentanglement term either fixed (semi-automated, SA) or learned (completely automated, CA). The reconstruction loss is split over multipole ranges in three configurations that sep
post-training - arxiv:2610.40409 · cs.AIRetrieving from Your Own Mistakes: On-Policy Training of Embedding Models for Multi-Hop Retrieval-Augmented GenerationChenjie Qiu, Zixuan Zeng, Shaobin Shi, Bingfeng Chen +2
Complex questions in retrieval-augmented generation require evidence spread over several documents, and an answer is reliable only when the retriever finds all of it. Chain embedding retrievers collect such evidence step by step, re-encoding the question together with the passages retrieved so far, and we find that they are already strong: trained only on gold histories, a small chain embedding retriever that calls no language model outperforms graph-based systems and is competitive with a search agent. This training, however, leaves a gap. At test time the retriever conditions on its own retrievals, which soon contain mistakes: by the third step, about half or more of the states it visits hold a wrong passage, and these passages lower its chance of retrieving a missing gold passage well beyond what the difficulty of the states explains. We propose on-policy state training, which trains the encoder on the states it reaches itself. At every epoch the current encoder retrieves chains for
retrieval-augmented - arxiv:2610.40401 · cs.ROContinuous-Action World Models: Learning Cross-Embodiment Spatiotemporal Dynamics PriorsLiangyu Li, Qingwen Liu
We introduce Continuous-Action World Models (CoAWM), a framework for learning cross-embodiment spatiotemporal dynamics priors that support robot action generation across embodiments and execution frequencies. At its core is a shared latent function space that captures the spatial and temporal structure of robot actions. We learn this space through two stages. First, function pretraining brings trajectories from different embodiments and sampling frequencies into a common representation, learning shared dynamics priors through action reconstruction and consistency across temporal sampling views. By accommodating heterogeneous robot trajectories, this stage broadens the range of data sources available for pretraining and facilitates scaling to larger and more diverse datasets. Second, we connect the pretrained function space to a world-model backbone, enabling observations, language instructions, and the current robot state to condition the prediction of a latent action function. A singl
world model - arxiv:2610.40394 · cs.AIScaling Mathematical Proof Supervision from Structured TextbooksYouqing Fang, Zhouqi Hua, Yinhao Tang, Yuzhe Gu +6
Mathematical supervision for language models is typically flattened into isolated problem-solution pairs, discarding the curricular organization and evidential structure of the original sources. We introduce TOME-Math (Textbook-Oriented Mathematical Evidence), a large-scale corpus that converts 2,118 textbooks spanning the core areas of university-level mathematics subjects into approximately 380K structured proof-supervision records while preserving chapter order, proposition-proof links, dependency references, and source provenance. Leveraging these structural signals, we explore curriculum learning across multiple mathematical subjects and characterize scaling patterns with supervision volume and source breadth. At full scale, fine-tuning on TOME-Math raises the unweighted mean score across our held-out evaluations and established public benchmarks by 16.7 percentage points over the base model, while exceeding all three public instruction-data baselines evaluated in aggregate. Our f
curriculum learningbenchmark - arxiv:2610.40379 · cs.CLAgentic LLM reasoning in pediatric diabetes outcome prediction from multimodal medical recordsDaniel Palacios, Guantong Qi, Anna Kadi, Natali Borrego +7
Large language model (LLM) agents are proposed for clinical prediction but are untested in pediatrics, where cohorts are small and outcomes are uncommon. Using records from 2,489 youth with type 2 diabetes (T2D) and 4,018 with type 1 diabetes (T1D), spanning structured data, note-derived social determinants and, in T1D, continuous glucose monitoring data, we predicted seven T2D outcomes, four also in T1D, at 24 months from 18 months of follow-up. CEDAR (Clinical Evidence-Driven Agent for Risk-prediction), an evidence-citing, self-auditing LLM pipeline, outperformed trained machine learning in both cohorts (pooled area under the receiver operating characteristic curve 0.815 versus 0.656 to 0.700 in T2D and 0.756 versus 0.603 to 0.631 in T1D), most on the sparsely screened outcomes. Only 4 of 70,246 values it cited did not match the chart. Reading only the evidence and reasoning cited by CEDAR, two pediatric endocrinologists reached the same prediction as CEDAR in about nine of ten cases
agentagentic - arxiv:2610.40373 · cs.ROHiFOR: Hierarchical Foresight with Action-Guided World Grounding for Vision-Language-Action ModelRongxing Ding, Hongyu Qu, Kehong Wu, Xudong Wang +3
Vision-Language-Action (VLA) models have demonstrated promising capabilities in robotic manipulation by integrating visual perception, language understanding, and action generation. Recent advances have further incorporated future prediction to enhance the understanding of environmental dynamics and task progression. However, existing methods often model short-term dynamics and long-term task objectives separately, while directly incorporating predicted futures into action policies without sufficiently considering their relevance to current action requirements. To address these limitations, we propose HiFOR, a hierarchical foresight framework with action-guided world retrieval. Specifically, Hierarchical Foresight Prediction jointly predicts short-term state evolution, semantic milestones, and final goal states through a Hierarchical Dependency Mask (HDM), establishing dependencies across temporal horizons. Building upon these predictions, Structure-Aware Latent World Modeling organize
vision-language-actionmanipulationworld modelbenchmark - arxiv:2610.40370 · cs.ROThe Athlete as a Scientist of One: Agentic Bayesian Experimental Design for Individualized Physiology from Consumer WearablesBehnam Moradi
Large language model (LLM) agents for personal health read wearable data and recommend, but do not learn about the individual in a way that can be audited or falsified. We present EpiCoach, an agentic system that treats one athlete, a cross-country mountain biker, as the subject of accountable N-of-1 science. Claims about the athlete live in an evidence-gated hypothesis registry: each must issue a probabilistic prediction about a future session before it occurs, and its belief moves only through computed evidence from a logged statistical fit. Predictions come from a mechanistic digital twin inferred from heart rate, GPS and barometric altitude: a lagged heart-rate response model, a fitness-fatigue model with learned time constants, and a latent technical-descending skill state, measured on discovered trail segments. The agent prescribes the next training week as a constrained experiment maximizing information gain subject to load and safety constraints, around two instrument rides. We
agentagenticbenchmark - arxiv:2610.40364 · cs.ROIs It the Curriculum or Just the Easy Start? Benchmarking Unsupervised Environment Design on QuadrupedsJakub Gilewicz, Dominik Galus, Piotr Wojdas
Unsupervised environment design (UED) methods such as PLR, ACCEL, SFL and PAIRED are compared mostly on 2D mazes, grid worlds and BipedalWalker, while terrain curricula for legged robots are hand-designed and evaluated against a few baselines. We present a benchmark that runs fourteen UED, domain-randomisation and robotics curricula under one protocol on five 3D quadrupeds in MuJoCo MJX: the closed-chain W01-TEK and the open-chain Unitree Go1 and Go2, Google Barkour vB and Boston Dynamics Spot. Levels are vectors of eight interpretable terrain parameters, and all curricula drive the same PPO student at matched environment-step budgets. The evaluation reports success on random and held-out structured terrains, interquartile means with stratified bootstrap intervals, capability envelopes and a level-by-level domination comparison. We compare methods within each robot and report, per level set, the top group of methods not significantly below the best. No method is in the top group on eve
quadrupedbenchmark - arxiv:2610.40356 · cs.ROJEDI: Joint-Embedding Demonstration Compression for In-Context Imitation LearningJiazhao Liang, Zhewen He, Shuaihang Yuan, Hao Huang +1
In-context imitation learning (ICIL) adapts robot policies through test-time demonstrations, but long multimodal trajectories constrain how many examples fit in context. We ask what a compressed demonstration should preserve. We propose JEDI, which compresses demonstrations into latent slots trained to predict the query episode's next-stage keyframe in a frozen video-JEPA space. Task and process counterfactuals help the slots preserve execution details, while the predictor selects demonstrations from a bank. Post-training with keypoint tokens enables human-hand conditioning on robots. Experiments on LIBERO and real robots show that JEDI supports many-shot context; LIBERO objective controls show improved success and demonstration following over action-only compression. These findings support query-future prediction as a training objective for preserving execution-relevant information in compressed demonstrations.
liberopost-training - arxiv:2610.40350 · cs.CLDon't Lose the Thread: Empowering Long-Horizon LLM Agents with Cognitive Resource Self-AllocationZheng Zhu, Xiaozhuan Liang, Xi Chen, Bei Yu +1
Agents powered by large language models (LLMs) have demonstrated remarkable progress in solving complex reasoning tasks. However, LLM agents often falter on long-horizon tasks due to cognitive overload, as their working memory becomes cluttered with expanding and irrelevant information, which dilutes their attention and hinders effective planning and reasoning. To mitigate this challenge, we introduce Cognitive Resource Self-Allocation (CORAL), a novel reasoning paradigm that empowers agents to proactively optimize their context. Implemented as an agent-callable working memory management toolset, CORAL allows an agent to checkpoint its task progress and verified facts, and adaptively initiate a new problem-solving episode by resetting cluttered context while preserving only the checkpointed information, enabling it to resume reasoning from a clean state. We further enhance the agent's checkpointing capabilities using a Multi-episode Agentic Reinforced Policy Optimization algorithm. On
memoryagentllm agentagenticbenchmark - arxiv:2610.40349 · cs.ROTouchBridge: Tactile-Conditioned Human-to-Robot Transfer for Dexterous ManipulationHaozhe Zhang, Gu Zhang, Zirui Zou, Yuzheng Chen +6
Human demonstrations offer an efficient and scalable source of data for dexterous robot manipulation, but differences in sensing and hand morphology complicate the transfer of contact-rich interactions. We present TouchBridge, a hardware--software co-designed framework for human-to-robot data transformation. A wearable human demonstration interface and a corresponding robot interface establish cross-embodiment correspondence in visual and fingertip tactile sensing. We introduce Tac-Retarget, a tactile-conditioned retargeting method that uses measured contact to align fingertip contact patches and preserve their spatial relations within and across hands. Unlike object-aware or simulation-based transfer pipelines, Tac-Retarget requires no object mesh, object-pose estimate, or physics simulation for retargeting, enabling direct and scalable conversion of human demonstrations into robot-domain training data. Combined with visual synthesis and camera-anchored action conversion, TouchBridge
vision-language-actionmanipulationdexteroustactile - arxiv:2610.40345 · cs.LGA Unified Hessian-Guided Framework for Post-Training QuantizationMads Toftrup, Marco Federici, Boris van Breugel, Paul Whatmough +1
Low-bit quantization enables efficient deployment of large language models by reducing memory footprint and computational cost. Achieving high accuracy at low precision requires selecting quantization ranges, linear transformations, and weight rounding decisions. Existing approaches typically optimize these components independently using local reconstruction objectives, which fail to account for the varying impact of quantization errors on downstream predictions. Direct optimization of task-level metrics would provide a more faithful objective, but is challenging due to the discontinuous nature of quantization operations and the high computational cost of evaluating their downstream effects. In this work, we derive a Hessian-guided approximation of the output change induced by joint weight and activation quantization. Based on a tractable Fisher approximation, we formulate a unified post-training quantization framework that enables the joint optimization of quantization ranges, linear
memorypost-training - arxiv:2610.40339 · cs.CVWizard: Manipulate Any World in Real Time via Agentic ControlYuzhe Wu, Zijun Lin, Yeying Jin
Recent advances in video world models have enabled the generation of increasingly realistic and interactive virtual environments, opening new possibilities for player-customizable gameplay. Agentic frameworks further advance this direction by decoupling explicit game logic and controllable world representations from visual generation. However, existing approaches support only simple game settings and predefined motion patterns, generalize poorly to unseen environments and characters, and offer limited real-time interactivity. In this paper, we introduce Wizard, an agentic game world model that enables real-time, precise control over character motions while generalizing zero-shot across novel scenes, characters, and actions. To achieve this, we construct a diverse dataset spanning 750 scenes and 178 characters, incorporating skeleton and geometry information as structured representations. We leverage these representations as intermediate proxies to condition a video generation model, to
world modelagentic - arxiv:2610.40333 · cs.CVRFW-A: Agentic Procedural Generation of 400K Articulated Assets over 500 CategoriesYuanbo Lyu, Changjian Pan, Saining Zhang, Yi Yang +18
High-quality articulated assets are essential for embodied perception, simulation, and manipulation, yet constructing them at scale remains difficult. Manual annotation and procedural modeling require substantial human effort. Recent agentic systems automate asset construction through program synthesis, but repeat costly reasoning and refinement for every new asset. Our key observation is that assets within a category can share construction logic despite variations in structure and appearance. Agentic synthesis and procedural generation are therefore complementary: agents can construct reusable programs, while procedural execution reuses the resulting logic across many samples. We therefore present RFW-A, which combines the two in a new articulated-asset generation paradigm. Starting from existing articulated programs, RFW-A organizes reusable components and their dependencies into category-level generators. Agents test and refine these generators across sampled configurations and moti
embodiedmanipulationagentic - arxiv:2610.40327 · cs.CVRadiology Guideline-derived Decision Graph for Efficient Clinical Decision SupportXuewen Zhou, Zhen Chen
Radiological risk stratification and management planning are essential downstream tasks in medical imaging. Automating these tasks is challenging because clinical guideline criteria are distributed across prose, tables, and cross-referenced definitions. We present Guideline-derived Decision Graph (GDGraph), a framework that decouples guideline compilation from case evaluation. During compilation, a large language model partitions guidelines into section-based chunks, extracts local decision subgraphs with linked definitions, and aggregates related branches into an executable decision graph. During inference, a fast System One model answers localized clinical questions while deterministic traversal resolves Boolean logic, turning opaque single-pass predictions into a traceable decision path without task-specific training. Across 347 cases spanning renal cyst, adnexal mass, and thyroid nodule guidelines, GDGraph improves System One accuracy from 57.3% to 90.8%, outperforming retrieval-au
retrieval-augmented - arxiv:2610.40318 · cs.AIDecision and System One models in rheumatology: accuracy and calibration on specialty-training examination questions and diagnostic case vignettesAlfredo Madrid-García, Diego Benavent, Fabian Lechner, Phillip Kremer +2
Objectives: To evaluate the accuracy and calibration of Decision and System One models on rheumatology examination questions and diagnostic vignettes. Methods: Two benchmarks were analysed: 172 rheumatology questions from Spanish specialty-training examinations (expanded RheumaMIR dataset) in Spanish and English, and 70 English diagnostic vignettes from RheumBench with 5-18 candidate diagnoses, 52 of which included the reference diagnosis. Four System One models (Jev 1.13, Clef, Microsoft-Decision-1, Laya) and GPT-6 Luna (OpenAI Decisions API) were compared with GPT-6 Sol and two local generative models. The primary outcome was accuracy; secondary outcomes included calibration, discrimination, accuracy of answers with probability >=0.90 and language differences. Results: In RheumaMIR, Jev's accuracy was 93.0% in Spanish and 91.9% in English, below GPT-6 Sol (99.4%; Holm-adjusted P<=0.005); GPT-6 Luna reached 97.7%, Clef 91.9-93.0% and Microsoft-Decision-1 86.0-87.8%, without detectable
benchmark - arxiv:2610.40305 · cs.LGRimuru: Unified Memory Management for Modern LLM Serving via Elastic Expert OffloadingZhiyuan Fang, Dingyang Li, Jianfeng Zhong, Bin Wang +5
Modern LLMs strain device memory through large model weights and attention states, particularly in resource-constrained scenarios. Existing systems separately provision memory for model weights and attention states, imposing fixed fixed boundaries that cannot adapt to the changing memory demands of workloads. Under heavy workloads, insufficient state capacity limits request concurrency and reduces throughput; under light workloads, idle memory in one region cannot be reallocated to another, wasting device memory. Meanwhile, modern LLM architectures introduce heterogeneous tensors with diverse layouts and lifetimes, further complicating memory management. These problems motivate unified memory management for heterogeneous tensors. However, a shared pool alone does not make memory directly reusable across tensors with different allocation granularities and execution layouts. This paper presents Rimuru, a modern LLM serving system that unifies memory management through elastic expert offl
memory - arxiv:2610.40293 · cs.MAMAServeBench: Benchmarking Multi-Agent LLM Serving Systems with Hierarchical Collaboration WorkloadsShengxuan Qiu, Haochen Huang, Shuzhang Zhong, Zizhuo Fu +3
Large language model (LLM)-based multi-agent systems (MAS) coordinate specialized agents to solve complex tasks, turning each user request into interdependent model and tool operations. Unlike independent requests, these operations become ready at different times and consume evolving context as agents coordinate and exchange results. This creates changing patterns of concurrency, computation, and memory demand that can expose different bottlenecks in LLM serving hardware. Yet benchmarking these effects requires capturing the collaboration rules that generate serving workloads, beyond the requests observed during execution. We present MAServeBench, a benchmark that treats multi-agent collaboration as a configurable dimension of serving workloads for LLM serving and hardware evaluation. Its composable stages capture how agents coordinate and exchange context in documented MAS workflows, enabling native executions to generate different workloads even when task inputs and tools remain unch
memorymulti-agentagent systembenchmark - arxiv:2610.40292 · cs.LGDolphinRouting: Cluster-Level Binary Routing for Disaggregated LLM Inference ServingYishuo Yuan, Shaoyu Yang, Chunrong Fang
Prefill-decode (PD) disaggregation isolates the two phases of large language model serving, but request-level path selection must account for prefill work, cache reuse, and decode demand. We present DolphinRouting, which learns this choice from end-to-end SLO feedback under a given P/D resource allocation. Requests are clustered by raw prompt length, cache-adjusted input lengths on both worker pools, and output-length information. Offline policy search learns one binary action per cluster. Online feature extraction, cluster assignment, and table lookup select local or remote prefill. We evaluate DolphinRouting on a real PD-disaggregated cluster using three public benchmarks, a mixed workload, and real data. Across the mixed workload and real data, DolphinRouting with $k$-means reduces P90 TTFT by 5.7--7.9\% and P90 TPOT by 2.6--10.1\% relative to Dynamo.
benchmark - arxiv:2610.40287 · cs.LGOnly Accepted Tokens Count: An Energy Accounting of Speculative Decoding Under Latency TargetsBole Ma, Ayesha Afzal, Jan Eitzinger, Harald Koestler +1
Large language model serving is limited by facility power. Energy per token determines how many tokens a datacenter can serve, while a latency target bounds how quickly each user receives them. During decode, every weight is read from memory for every generated token. Batching amortizes this read across users but makes no individual user faster. Speculative decoding does: a drafter proposes several future tokens, and the target model verifies them in a single forward pass that reuses the same weight read. Because every drafted token must be verified whether or not it is accepted, the energy speculation saves depends on how many drafted tokens are accepted. This saving is hard to pin down. It varies across target models from different attention families and across drafting schemes, and it depends on the text being served. Every drafter we evaluate accepts fewer tokens on conversation than on code, and every trained drafter accepts fewer on Romansh, a low-resource language, than on Engli
memory - arxiv:2610.40285 · cs.ROMUSBRIC: LEGO-Compatible Pneumatic Muscle Bricks for Reconfigurable Musculoskeletal RobotsQiyao Wang, Yinglei Zhu, Liang Zhou, Wenzhe Gong +2
Musculoskeletal robots combine compliant actuators with articulated skeletons, but exploring alternative configurations often requires repeated fabrication. We present MUSBRIC, a modular hardware platform and accompanying MuJoCo-based simulator for reconfigurable musculoskeletal robots. LEGO-compatible endcaps turn fiber-reinforced extensile pneumatic artificial muscles into reusable, plug-and-play muscle bricks that connect directly to standard LEGO components. Experimentally identified actuator behavior supports corresponding musculoskeletal simulations. An elbow-joint prototype validates load-dependent flexion and extension, showing that pulley-guided routing enables reversible motion and a mechanical stop stabilizes the straight posture. A robotic hand demonstrates compliant grasping and multi-joint construction. Matched hardware and simulation sweeps across 15 leg configurations demonstrate systematic exploration of attachment geometry, while a quadruped assembled from four legs d
quadrupedgrasp - arxiv:2610.40283 · cs.LGLatent Path Exploration via Potential-Guided Subtrajectory BalanceXiaoling Zeng, Qi Zhang
Continuous latent reasoning enables language models to perform intermediate computation without generating explicit reasoning traces. Advances in stochastic latent reasoning allow multiple paths to be sampled in continuous space, but stochasticity alone does not ensure effective exploration. A key challenge is guiding stochastic transitions between intermediate latent states toward useful reasoning paths. We introduce SubLAT, which couples a relative answer potential with Subtrajectory Balance to learn stochastic latent exploration. The relative potential evaluates the answer support of sampled states against replacement states, with a correctness signal at the terminal state. Subtrajectory Balance relates potential differences to forward and backward transition densities across subtrajectories of different lengths. The model is jointly trained with answer supervision, CoT supervision, and representation distillation. Experiments on arithmetic and deductive reasoning show increasing so
benchmark - arxiv:2610.40276 · cs.CVTaskFrame: Connecting Task-Level Planning with Frame-Level Control for Game AgentsGuangzhi Wang, Junfu Pu, Yiran Wang, Xingyilang Yin +9
Building general game agents requires both high-level task reasoning and low-latency action control in rapidly changing visual environments. These capabilities favor models with substantially different computational profiles, making them difficult to realize within a single model. In this work, we present TaskFrame, a hierarchical planner-controller framework that connects task-level planning with frame-level keyboard-and-mouse control. A vision-language task planner decomposes a high-level goal into an ordered sequence of atomic instructions and is optimized through supervised learning and dependency-aware GRPO. The hidden states of special goal tokens that terminate each atomic instruction provide continuous goal embeddings that are cached and reused across control steps. A causal Transformer controller combines the embedding of the active goal with recent observations to predict discrete keyboard and mouse-button actions and continuous camera motion at every game step. We train the
agent - arxiv:2610.40256 · cs.CLWhen Requirements Interact: Scaling Long-Horizon Coding via Interdependent SpecificationsYishuo Cai, Minzheng Wang, Liang Zhao, Linghao Zhang +4
To succeed in real-world long-horizon software development, coding agents must satisfy multi-dimensional task specifications, such as interface compatibility and error handling, and do so without breaking existing dependencies. However, current works lack specification in both quantity and diversity, often emphasizing primary goals while ignoring complex constraints. Synthesizing tasks to bridge this gap presents two challenges: naively adding specifications risks reducing tasks to trivial checklists rather than compelling joint reasoning, and verifying these subtle interactions risks overlooking hidden flaws or rejecting valid alternatives. To address these challenges, we introduce CodeSpecWeave, a framework for synthesizing tasks with rich, interdependent specifications. Grounded in real developer seeds, a synthesis agent iteratively develops both coupled task requirements and executable tests, while a review agent audits solution trajectories from multiple models to guarantee test r
agentbenchmark - arxiv:2610.40254 · cs.LGStatistical Analysis of Synthetic Data with Verification: A Linear Regression FrameworkHeejune Sheen, Harrison H. Zhou
Synthetic data are widely used for training modern AI systems. However, their statistical understanding remains limited. It is well recognized in statistics that repeated self-training on synthetic data can amplify errors inherited from pre-trained models, leading to model collapse. Verification plays a crucial role in selecting high-quality synthetic data to help prevent model collapse and improve model performance. We study verified self-training in a linear regression framework, where a pre-trained estimator generates synthetic data with injected generation noise and a verifier selects samples according to a verification threshold. Post-training on the accepted samples updates either all coordinates (full fine-tuning) or only an active subset (partial fine-tuning). Our analysis reveals three potentially counterintuitive effects. We establish that tighter verification does not always lead to lower prediction risk. We also find that a suitable level of injected generation noise can re
post-training - arxiv:2610.40250 · cs.AIDistributional Reward Calibration and Budget-Neutral Credit Assignment for Multimodal RLVRLang Chen, Zongxin Xu, Borui Wang, Anran Liu +1
We revisit advantage estimation in Group Relative Policy Optimization (GRPO) and observe that any token-level advantage admits a two-stage factorization: a \emph{sample-level baselining} step that converts scalar verifier rewards into a comparable sequence-level signal, and a \emph{token-level credit assignment} step that distributes that signal over the trajectory. Existing estimators typically address only one of these stages: affine baselines preserve the shape of each task's reward law, so heavy tails and binary cliffs propagate into the gradient, while token-level credit rules tie the size of each token's update to raw confidence scales and need not conserve the total credit issued by the verifier. We propose \textbf{TACE} (Two-stage Advantage and Credit Estimation), a two-stage estimator that addresses both stages. At stage~(S1), we replace linear standardization with a one-dimensional optimal-transport map to a standard normal, which gives a common population reward law across t
benchmark - arxiv:2610.40248 · cs.ROLLM-guided Planar Loco-Manipulation for Quadruped RobotsGeorges Jetti, Michael Khayyat, Jean-Pierre Sleiman, Stefano Arrigoni +1
Robotic navigation typically assumes a collision-free route to the goal exists; in cluttered environments with movable obstacles, however, reaching the goal may require physical interaction. Classical solutions encode these behaviors through geometric rules and task-specific logic, while learning-based approaches often constrain the decision space to low-dimensional motion commands or predefined action sets. We instead consider a broader decision space, jointly selecting the skill, relevant object, and geometric target. Our language-conditioned task planner, fine-tuned with LoRA on synthetic data, maps natural-language tasks and a structured scene representation to variable-length sequences of parameterized actions, executed by low-level reinforcement-learning policies. The planner learns the geometric relations underlying interactive navigation and outperforms a classical planner and a prompted frontier model on interactive navigation tasks. Comparisons with models trained from scratc
manipulationquadruped - arxiv:2610.40246 · cs.AIEnvOS: Auditing Computer-Use Agents When the World Moves On : Phenomenon-Based Task Construction, Certification, and Evidence-Based EvaluationAnkush Kadu, Athira Gopal, Niharika Singh, Arpit Jain +1
Computer-use agents built on large language models (LLMs) work in software that can change while they act: an appointment moves, a policy is revised, a record is updated. Testing agents under such changes is harder than it looks. When an agent fails, the cause may be the agent, but it may also be a change that did not fire as intended, an update the agent never received, or a grader that scored the episode wrongly. We introduce EnvOS, a framework for building and certifying environments that change mid-task. Before any agent runs, EnvOS checks that each change fires correctly, that the task remains solvable, and that the grader separates correct from incorrect behavior; after each run, separate code re-checks the verdict from archived evidence. We apply it to stale information, where a value the agent has already seen changes before the agent acts on it. In clinic booking, Claude Opus 5 succeeds on 0/16 tasks when nothing extra is shown, 2/16 when reminded to re-check the source, and 1
agent - arxiv:2610.40238 · cs.CLScienceJudge: Pattern Perturbation Augmentation for Sentence-Level Scientific Claim VerificationChuhai Cai, Xin Chen, Min Yang, Feng Jiang
Sentence-level scientific claim verification determines whether each sentence in an AI-generated scientific paragraph is supported by a given multimodal evidence bundle. Fine-grained reporting errors often preserve fluent and plausible statements while introducing subtle evidence-claim deviations, making them difficult to detect under sparse and imbalanced supervision. We propose Pattern Perturbation Augmentation (PPA), which models reporting errors as label-specific faithfulness breaking transformations and induces reusable error patterns from labeled examples. PPA applies these patterns to supported claims to construct controlled hard negatives and expands their coverage through pattern-level, operation-level, and cross-domain augmentation. Based on these augmented views, we develop ScienceJudge, which integrates PPA-trained vision-language predictors and instruction-guided LLM predictors through class-weighted ensemble inference. On the local development set, ScienceJudge improves t
leaderboard - arxiv:2610.40227 · cs.AIEpiAgent: Epistemic Self-Play for LLM AgentsXuanfa Jin, Xinyu Cui, Zhijian Ma, Xue Yan +3
Epistemic game theory explains an agent's strategic behavior via the beliefs it holds about others and how these beliefs are updated as the interactions unfold. Modern LLM-based agents, however, often lack an explicit mechanism for learning and updating such beliefs to guide decisions under incomplete information. To bridge this gap, we propose EpiAgent, a framework that equips LLM-based agents with an explicit, learnable epistemic layer, which maintains beliefs over others' private information and combines these beliefs with the agent's available information into an epistemic state for action selection. Concretely, the agent's policy is decomposed into two coupled stages: Epistemic Update, which updates these beliefs from interaction histories, and Action Planning, which selects actions conditioned on the updated epistemic state. By formulating their dependency as a bi-level optimization problem, we introduce Epistemic Self-Play, an algorithm that alternately optimizes the two stages
llm agentmulti-agentself-play - arxiv:2610.40212 · cs.CVBenchmarking color judgment in multimodal LLMs with CIE-computable ground truthShuwei Yue, Kaida Xiao, Zhihua Wang, Qiao Chen
Multimodal large language models (MLLMs) are now used as zero-shot judges of image quality, but their color judgments are rarely tested with colorimetric ground truth. We present a five-task benchmark based on the CIE colorimetric framework: AWB pair judgment, skin-tone bias direction, CCT estimation, CIE Delta E 2000 pair ordering, and color naming under changing illuminants. We test eight MLLM endpoints, including four Qwen3-VL variants, GLM-4.5V, GPT-5.4, Claude-Opus-4.7, and Gemini-3.1-Pro. The models often solve coarse binary AWB judgments, but they remain weak on numeric CCT estimation and on some fine color-difference judgments. On ColorChecker-24 stimuli, the three closed-weight frontier models reach CCT MAE of 715-1056 K, compared with 1358-2133 K for the five open-weight models. A non-ColorChecker control shows that this advantage is partly template-dependent. The results suggest that current MLLMs have useful color words and coarse visual priors, but their color judgments ar
benchmark - arxiv:2610.40210 · cs.ROAcrossDEX 1.0: From Human Motion Pretraining to Bimanual Dexterous Manipulation across Hand EmbodimentsXiaoyu Yang, Nan Wu
Bimanual dexterous manipulation requires continuous coordination of wrist motion, finger contact, and object motion. Human videos provide rich manipulation experience. Using this experience for robot control requires aligning human poses with robot commands and accounting for embodiment-specific kinematics and contact dynamics. We present AcrossDEX 1.0, a learning system that connects human motion pretraining, simulation-based expert learning, robot demonstration collection, and visual closed-loop execution. Pretraining on HOT3D Aria predicts one second of bimanual motion from five images and thirty annotated hand states over the preceding second. Robot adaptation preserves the visual encoder, temporal backbone, and wrist parameters. Shared wrist representations, separate finger fields, and validity masks support both six-channel Inspire hands and sixteen-joint Allegro hands. The robot policy predicts sixty control targets from frontal images, command history, and task conditions. It e
manipulationdexterousrobot policy - arxiv:2610.40207 · cs.RORoboAtom: Large-scale Atomic-Action Supervision for Robot LearningXueyuan Wei, Xuanxu Wang, Xudong Li, Yuping He +9
Fine-grained action annotations can support skill reuse and compositional generalization in robot policies. However, robot datasets typically provide task-level instructions without subtask annotations. Although some include subtask labels, inconsistent formats and conventions hinder cross-source joint training. We propose RoboAtom-Data, which adds unified, time-aligned atomic-action annotations to existing data from multiple sources, covering 1.45 million trajectories, 6.88 million skill-labeled segments, and 16,882.1 hours. Our pipeline combines a unified protocol, large-scale manual annotation, propagation across repeated executions with the same ordered skill sequence, and multi-stage quality control. We evaluate the dataset through automatic atomic-action annotation and robot control. We construct RoboAtom-Bench to standardize evaluation of skill recognition and temporal boundary localization. Trained on the annotated data, RoboAtom-Anno achieves better automatic annotation perfor
liberorobotwinbenchmark - arxiv:2610.40192 · cs.AIBridging Semantics and Structure: A Theoretically Grounded Framework for Link Prediction on TAGsChen Shao, Ruirui Wang, Tobias Kafer, Danai Koutra
Link prediction on text-attributed graphs (TAGs) is a fundamental problem with applications in knowledge discovery, recommendation and beyond. Recent progress has highlighted the promise of combining Graph Neural Networks (GNNs) with Language Models (LMs) in order to leverage the complementary structural and semantic signals. However, due to the distinct inductive biases of GNNs and LMs, the design of an optimal fusion strategy remains an open research question. In this work, we present Bridge-LP, an intuitive, simple, efficient and theoretically grounded framework that unifies graph and language representations. It is built around three theoretically-justified design components: (D1) jumping connection of textual proximity, which enables robust LM training; (D2) compressed embedding, which encodes neighborhood information long exploited by classical link prediction heuristics, thereby enriching structural signals in a learnable way; and (D3) cache embedding training strategy, which re
benchmark - arxiv:2610.40191 · cs.CVVideo Generation Meets Agent Harness: Agentic Video Generation, Agentic World Modeling and Recursive Self-ImprovementLiaoyuan Fan, Yuhao Zheng, Dongxing Mao, Hangyu Ran +17
Video generation is shifting from end-to-end synthesis toward agentic generation, where planning, tool use, and feedback enable coherent, controllable multi-shot videos. In parallel, video world models provide action-conditioned observations and predicted futures, serving as interactive environments for agents. We study their intersection through the agent harness: an executable runtime that constructs context, exposes tools, maintains state, routes feedback, and manages budgets around video models. This report examines three directions. First, agentic video generation, where harnesses coordinate scripting, shot planning, asset management, production, and editing, alongside the evaluation challenges of these systems. Second, agentic world modeling, where harnesses mediate observation, control, execution, and state. We discuss hybrid designs in which game engines or agent-written programs maintain world state while video models render observations, and how agents and world models can mu
world modelaction-conditionedagentmulti-agentagentictool use - arxiv:2610.40190 · cs.CVAdMuse: Skill-Guided Creative Advertising Image GenerationMinyan Luo, Yuxin Zhang, Meng Zhao, Xincan Wang +5
An outstanding creative advertising image should both attract attention and communicate a product's positioning. How to choose a persuasive conceptual bridge before rendering begins is challenging for advertising generation. We present AdMuse, a skill-guided multi-agent framework that distills high-quality advertisements into reusable creative skills for cross-domain advertising strategy. Rather than retrieving isolated exemplars, AdMuse abstracts recurring bridge patterns into a skill library that can guide new product briefs. At generation time, AdMuse coordinates skill routing, multi-candidate ideation and selection, and VLM-guided visual refinement: diverse skills propose alternative concepts, relative judgment selects the most promising direction, and vision-language feedback helps realize the selected intent in the final image. This design treats advertising creativity as a structured decision process rather than a single prompt-writing step. Experiments across diverse product an
multi-agentagent framework - arxiv:2610.40186 · cs.AILocating and Closing Information Gaps in Multi-Agent SystemsYaxin Gong, Chongming Gao, Gangyi Zhang, Chenxiao Fan +6
Multi-step Web tasks often require several LLM agents to gather and combine evidence. The communication graph of a fixed workflow restricts which agents can read each other's outputs: evidence retrieved by one agent may never reach the agent that needs it, leaving an information gap and an incomplete or incorrect answer. Adding a connection can deliver the missing evidence, but can also lead the receiver to adopt an incorrect upstream conclusion that displaces a correct answer; the final output alone does not separate the two. We propose Selective Probing and Link Insertion for Closing information gaps in multi-agent Execution (SPLICE), which locates information gaps per query and decides what to adopt after adding an edge. After the original workflow runs, SPLICE uses the query and node outputs to locate a possible gap, adds one directed edge from a source to a receiver to deliver the relevant evidence, and re-executes the receiver and affected downstream nodes, producing a candidate
agentllm agentmulti-agentagent systemtool usebenchmark - arxiv:2610.40175 · cs.CVBeyond CCT: Multi-channel SPD recovery for per-shot camera color correctionShuwei Yue, Hongke Wei, Qiao Chen
Modern cameras correct color via two or three pre-calibrated 3x3 matrices interpolated by Correlated Color Temperature (CCT). This collapses the illuminant SPD to a scalar. LEDs, fluorescents, and mixed sources sit off the Planckian locus, and metameric illuminants with the same chromaticity demand different optimal CCMs that a CCT pipeline cannot distinguish. We recover the illuminant SPD instead using non-negative least squares against a basis dictionary, conditioned on the camera white point and a multi-channel ambient-light sensor (ALS). A closed-form 3x3 CCM is then fit per shot. On a synthetic benchmark of 300 illuminants and 3,100 reflectances, mean angular error drops from 1.89 to 1.20 degrees, with a 1.7x tail reduction on the LED-like off-Planckian group (n=45) where metameric ambiguity is tangible. The gain holds under realistic ALS imperfections and observation noise, and the same off-Planckian advantage carries over to 1,684 real measured SPDs.
benchmark - arxiv:2610.40166 · cs.CVEventSeeColor: Learning to Reconstruct Color from Monochrome EventsChuanzhi Xu, Langyi Chen, Haoxian Zhou, Zelin Zhang +4
Event cameras capture visual brightness changes under challenging lighting conditions. However, traditional event-based image reconstruction targets grayscale intensity and produces images without scene color. Color distinguishes surfaces with similar intensity. Reconstructing color from events can reveal appearance differences that reconstructed grayscale images cannot express. Although monochrome event cameras do not directly measure color, their event-triggering patterns depend on surface reflectance, illumination, and the sensor's spectral response. These patterns may contain cues that support color reconstruction. In this paper, we present EventSeeColor, which uses event-reconstructed grayscale and aligned events to jointly guide a diffusion colorizer. Grayscale-derived semantic context supports color inference, while region-balanced chroma supervision and local RGB refinement improve color distribution, brightness, and texture. In the combined evaluation on two real event dataset
event camera - arxiv:2610.40162 · cs.LGFisher Contrastive Learning to Mitigate Feature Suppression: A Case Study in Acute Promyelocytic Leukemia DiagnosisShushan Wu, Yongkai Chen, Xilin Gong, Yufang Liu +2
Self-supervised contrastive learning (SSCL) learns representations by contrasting augmented views of the same sample, but it can exhibit feature suppression, where dominant or easy-to-learn factors suppress other class-relevant signals, degrading the performance of representations, particularly in high-dimensional data with heterogeneous features. We introduce Fisher Contrastive Learning (FCL), which uses a regularized Fisher trace criterion to address feature suppression. FCL learns embeddings by maximizing the discriminative strength of Fisher discriminant functions, producing representations that preserve class-relevant information implied by the augmentation process. We prove that the representations learned by FCL could alleviate the feature suppression effects. We demonstrate FCL on simulated benchmarks designed to induce feature suppression and on a real-world biomedical application, pretraining on large white blood cell microscopy image dataset and transfer learning for diagnos
benchmark - arxiv:2610.40157 · physics.opticsCubic-Zn(1-x)Mn(x)Te optical modes across the phonon and phonon-polariton regimesM. Mina, A. Elmahjoubi, T. Ngulezhu, S. Thiam +17
Alloying offers a convenient route to tune the physical properties of semiconductors, at the price, however, of local lattice/bond distortions needed to accommodate the lattice mismatch. Being nearly lattice-matched and moreover highly resistant to lattice and bond distortions thanks to a largely covalent bonding, cubic Zn(1-x)Mn(x)Te sets a benchmark to test the sensitivity of lattice vibrations to minimum lattice and bond distortions induced by alloying. We explore experimentally the two transverse optical regimes of cubic Zn(1-x)Mn(x)Te (x<0.7), with ab initio calculations in support: (1) the phonon-polariton regime at quasi-infinite (photon-like) wavelengths averaging over the alloy disorder, probed by forward-Raman scattering, and (2) the purely-mechanical phonon regime down to minimum wavelengths nearly matching the lattice constant zooming into the alloy disorder, achieved by inelastic neutron scattering. The short Zn-Te bond is found to exhibit a compact optical duo distinguish
benchmark - arxiv:2610.40133 · cs.AIDynaRev: Benchmarking Web Agents under Dynamic Goal RevisionShiqi Zhang, Guochen Zhou, Zesheng Wei, Hai Zhao +1
Web agents are typically evaluated with goals fully specified before execution and fixed thereafter, whereas real users often revise their requirements during interaction. To study such interactions, we introduce DynaRev, a benchmark of 181 tasks to evaluate web agents under dynamic goal revision. DynaRev uses semantic milestones to deliver revisions at task-defined stages, covers additive revisions, unconditional overrides, and runtime-conditioned overrides, and evaluates correctness, responsiveness, and efficiency through trajectory-grounded evaluation. Five-model evaluations show that agents often satisfy only part of the task requirements, with performance varying across task sources, revision types, and models. Further analyses reveal varied milestone timing, mixed effects of dynamic delivery, and contrasting revision-handling outcomes, with gains from instruction rewriting while preserving browser state and losses from browser resets. DynaRev provides a testbed for studying how w
benchmark - arxiv:2610.40131 · cs.CVVideoMemBench: How Well Do MLLMs Remember in Video Conversations?Zhengdong Hu, Chao Wang, Kaituo Feng, Huanjin Yao +3
Multimodal large language models (MLLMs) can now process hour-long videos, making extended multi-turn dialogue over a single video a natural mode of use. Sustaining such a dialogue requires more than answering each question in isolation: the model must retrieve and reuse information established earlier in the conversation, both what it has seen and what it has said, an ability we term in-context video memory. Existing video QA benchmarks leave this ability unmeasured: their questions are mutually independent, never reference one another, and each is answerable from the video alone, so nothing ever requires carrying information across turns. We introduce VideoMemBench, in which each sample pairs a single video with a long multi-turn dialogue, whose turns are connected by typed cross-turn dependencies: an Episodic Dependency refers back to a visual episode described in an earlier turn, probing memory of what the model has seen, while an Answer Dependency builds on the model's own earlier
memorybenchmark - arxiv:2610.40124 · cs.CVIt Takes Two to Probe, Thousands to Retrieve: Putting Vision--Language Probes to the Retrieval TestIvana Beňová, Michal Gregor, Marián Šimko
Vision--language models (VLMs) are tested for fine-grained understanding with image--text matching probes, a model must prefer the correct image or caption over handcrafted negative. In retrieval, the correct item must beat thousands of alternatives. We measure how these two evaluation methods relate by using two probing benchmarks, \svo{} and SugarCrepe, for retrieval tasks. For 100 image and caption queries we judge the top-10 retrievals from full dataset of six VLMs, with human annotators and an LLM assessor. We than compare how well models rank pool of items on the same queries, while changing only where the negatives comes from: the benchmark, a random draw, or the judged-incorrect retrievals of other models. On \svo{} the accuracy of retrieving a positive over handcrafted negatives is significantly higher than for retrieved ones (the correct item is preferred in 85\% against 77\% of the time). Using pool of ten retrieved negatives leads to same accuracy as when one uses pool of 2
benchmark - arxiv:2610.40116 · eess.SYDual-Agent Reinforcement Learning for Vehicle Scheduling and Multi-Energy Dispatch in Electric-Hydrogen-Transport SystemsFulong Yao, Yiming Xu, Maurizio Albano, Liana Cipcigan +3
Electric-hydrogen-transport systems (EHTSs) require coordinated scheduling of electric and hydrogen vehicles together with multi-energy dispatch. These operations form a sequential decision process: upstream vehicle scheduling (VS) determines service allocation and order, thereby endogenously shaping the electricity and hydrogen demands faced by downstream energy dispatch (ED). This paper develops a dual-agent reinforcement learning framework tailored to the heterogeneous decision structures. The vehicle-scheduling agent combines proximal policy optimisation (PPO) with a Plackett-Luce policy to learn vehicle-priority permutations, which are then mapped to feasible charger and refueller assignments through a greedy allocation. The energy-dispatch agent employs twin delayed deep deterministic policy gradient (TD3) for continuous storage and energy-flow control. The directional coupling between the two decision layers is analysed, revealing that updates to the upstream scheduling policy c
agent - arxiv:2610.40115 · cs.CVSelfThinker: Self-Evolving Reasoning via Reliability-Verified Intrinsic Visual ThoughtsZhangquan Chen, Ruqi Huang
Unified multimodal models can both generate and understand images within a single set of weights, offering an opportunity to use image synthesis as training supervision for visual comprehension. However, existing visual chain-of-thought methods rely on human-annotated helper images during training and require slow image generation at test time. We present SelfThinker, the first self-evolving framework that turns a unified model's self-generated visual thoughts into stronger understanding without human image annotations. Given a visual problem, SelfThinker samples candidate helper images from its generation pathway and uses its frozen understanding pathway to score each candidate via conditional pointwise mutual information. High-utility visual thoughts then refine the generator via flow matching and serve as privileged teacher context for on-policy self-distillation. At inference, SelfThinker reasons directly from the original image in text without generating intermediate images. Altho
self-evolvingbenchmark - arxiv:2610.40109 · cs.ROCoupled Guidance for Flow MatchingMateo Clémente, Léo Maxime Brunswic, Rui Heng Yang, Amir Rasouli
Guidance methods are essential tools for steering the generation process in diffusion models, improving the generalizability of stochastic robotic controllers and enabling high-quality controlled image synthesis. Despite strong parallels between flow-matching and diffusion, integrating guidance into flow-matching models remains quite challenging. Recent work has proposed general guidance frameworks for flow-matching, however, they are direct adaptations of diffusion guidance methods and can prove unstable. In this work, we conduct a theoretical analysis of guidance in flow-matching and show that a naive translation of diffusion-based guidance to flow-matching breaks the coupling distribution used during training, resulting in weaker control over sampling accuracy. Building on this insight, we propose Coupled Guidance, a flow-matching-specific guidance mechanism derived from a geometric perspective that leverages the coupling structure inherent to flow-matching training. To validate the
manipulationdexterousbenchmark - arxiv:2610.40096 · cs.AIReward-Adaptive Co-Evolution of Policy and RubricYizhu Chen, Jianguang Jiang, Jiameng Huang, Chuyu Fang +4
Rubric-based rewards provide fine-grained, interpretable supervision for open-ended reinforcement learning, where holistic scalar rewards struggle to capture multidimensional response quality. Static rubrics can lose discriminative power as the policy distribution evolves, while existing policy--rubric co-evolution methods rely on globally prescribed update schedules that ignore heterogeneous reward informativeness. We introduce RACE, a Reward-Adaptive Co-Evolution framework. At the group level, per-rollout-group training-signal routing classifies groups as informative, diversity-collapsed, or rubric-failed, updating the policy only with informative groups while caching rubric failures. At the phase level, event-triggered phase switching enters rubric optimization upon persistent reward saturation and returns when rubric-reward improvement plateaus. Rubric phases train the generator on cached pairs using a frozen judge and a scale-normalized pairwise-margin reward. Across both levels,
evaluation protocol - arxiv:2610.40075 · cs.ROLearning Safe Flow Matching Robot Policies from Binary Feasibility OraclesJianming Ma, Rongjun Jin, Yizhi Chen, Yue Gao
Robot foundation models increasingly rely on flow-based action generation for general-purpose manipulation, yet the generated actions are not guaranteed to satisfy physical constraints such as joint limits and collision avoidance. This problem is especially challenging when the feasible set cannot be described analytically and is accessible only through black-box feasibility oracles. We study this practical oracle-only setting and propose an oracle-guided differentiable safeguard for safe flow-matching robot policies. The safeguard uses binary feasibility queries to correct unsafe updates in the forward pass, while extracting local boundary information to provide constraint-aware learning signals in the backward pass. Combined with discrete-time flow matching, the same safeguard is used during training and inference, enabling forward--backward aligned policy learning without explicit or differentiable constraint models. Experiments on obstacle-aware robotic manipulation demonstrate 100
manipulationrobot foundation model - arxiv:2610.40071 · cs.AIAgentic process evidenceYuval Kainan, Carmit Hershman
The way software gets built is changing dramatically. AI agents now create pull requests, select dependencies, refactor codebases, interpret test failures, and execute infrastructure changes across the entire Software Development Life Cycle (SDLC). AI agents are autonomous team members, but unlike human engineers, they are ephemeral - their reasoning, decisions, and tool calls vanish the moment the session ends. Modern Security incidents involving intent mismatch or injection are becoming increasingly common - forcing us to address AI Agents as an insider threat. No widely adopted standard yet enables agentic provenance and governance and its obvious that you can't hold an ephemeral process accountable. The git log was designed for human commits. It shows what changed. It does not show what an AI agent did, which model it used, which tools it called, or whether anyone was accountable for the outcome. Every governance framework, every audit trail, every compliance control in the SDLC wa
agentai agentagentic - arxiv:2610.40068 · cs.CVBRACHIO: Broad Representations Across Centres for Head-and-neck Imaging Oncology (HECKTOR 2026)Roxanne De Maeyer, Ine Dirks, Yannick Brackenier, Simon DeKeyser +2
The HECKTOR 2026 challenge benchmarks machine learning for head and neck cancer, where teams delineate primary and nodal tumour volumes, perform TN staging, and predict recurrence-free survival (RFS). This report describes the submissions for the HECKTOR challenge of team nucli. As the three tasks span both dense and global information from multicentric PET/CT scans, we utilise self-supervised pre-trained encoders, training a 2D DINOv2 encoder (PET-DINOv2-MIPs) on HECKTOR 2026's PET scans. Our final test submission ensembles a DynUNet trained from scratch with a SwinUNETR whose Swin trunk was initialised from a SimMIM-pre-trained checkpoint for segmentation, reaching a mean Dice of 0.5409. For TN staging, GTVp/GTVn geometric features from the segmentation masks are fit to stage probabilities by logistic regression (LR) and fused with PET-DINOv2-MIPs deep features (PCA + LR) via a second-level LR, reaching balanced accuracies of 0.3962 (T-stage) and 0.6432 (N-stage). For the RFS prognos
benchmark - arxiv:2610.40056 · cs.LGReinforcement Learning Under Runtime Confounding: Policy Learning and Off-Policy EvaluationRaphael Kim, Michele Santacatterina, Ramin Zabih, Keith Barnatchez
Reinforcement learning (RL) is increasingly used to train robust, adaptable systems across healthcare, e-commerce, and robotics. In these settings, legal, ethical, or logistical concerns may preclude certain key variables that drive decisions from being collected during deployment time. This gives rise to a setting known as runtime confounding: the logged data hold full observability, but deployment time lacks knowledge of certain variables, known as runtime confounders. Unlike single timepoint and non-adaptive settings, runtime confounding under reinforcement learning has the potential to propagate over time, as earlier decisions affect future states, observations, and action assignments, leading to biased learning. We address these problems with three contributions: we (1) develop policy learning algorithms under runtime-confounding, (2) develop semi-parametric efficient off-policy estimation procedures, and (3) provide diagnostics for understanding the price of runtime confounding.
policy evaluation - arxiv:2610.40036 · cs.AIHybrid AI Patterns: Thirteen Ways to Couple Language Models with Rules, and How to Choose Among ThemSiegfried Handschuh, Götz-Henrik Wiegand, Lorena Raichle
Language models handle text well, and rule engines give conclusions with proofs. Systems that combine the two are usually described by example, and the same coupling turns up under different names in agent frameworks, guardrail toolkits and neuro-symbolic research. This document defines thirteen patterns for coupling a language model with an executable rule language, in the catalogue form introduced for object-oriented design: intent, motivation, applicability, participants, collaborations, consequences, implementation, known uses and related patterns. Four patterns concern what crosses the boundary between model and rules, five who controls whom, three how the model takes part in the reasoning, and one how knowledge is kept over time. Five invariants hold for every combination, and five failure patterns name the couplings that break them. We relate the catalogue to published pattern catalogues and taxonomies of hybrid and neuro-symbolic systems and show where they overlap with it. A d
agentagent framework - arxiv:2610.40031 · cs.CLAI Cosmos: An Agentic Infra for Situated LearningAI Cosmos Team, Jinfeng Zhou, Yuxuan Chen, Fangming Chang +31
Large language models (LLMs) can improve task performance during learning, yet these gains do not necessarily reflect learners' ability to thereafter apply the knowledge and skills independently. We present AI Cosmos, an agentic infrastructure for situated learning that organizes learning around two complementary principles: scenario-based participation and cumulative assets of practice. AI Cosmos organizes learners and role-specialized agents within learning scenarios that integrate task-relevant knowledge and, when needed, executable workbenches. An asset-memory loop preserves the processes and outcomes of practice as reusable assets with explicit provenance and dependencies. We evaluate scenario-based interaction through two preregistered randomized controlled studies examining the acquisition of legal reasoning and programming competence, holding the same underlying LLM across conditions. AI Cosmos achieves greater learning gains than a conventional LLM in legal reasoning and highe
agentic - arxiv:2610.40030 · cs.LGCheap Retrieval and Observation Credit Assignment from One Quantity: The Value of InformationDaniel Strauss, Nikolaos Sotirakis, Zhenshan Bing, Alois Knoll
Memory in reinforcement learning raises two problems. Attention finds the relevant item by scoring the whole store, which is costly for robots under latency and energy limits. When a retrieved memory pays off, credit for acquiring it must cross the whole sequence. We argue that both are governed by one quantity, the value of information, and propose to estimate it from archived rollouts with a learned world model. We obtain that, when read at the observation, it yields a provable lower bound on the optimal value of the acquiring action, which floors its TD target and attributes this floor to the observation that earned it. Read at the decision, it defines the optimal retrieval set. A pilot retrieval system, to be trained and improved with this signal, solves a memory task with reduced context.
world modelmemory - arxiv:2610.40026 · cs.RORoboSeed: Multidimensional Evaluation of Vision-Language-Action and World-Action Models: Capability, Resource Affordability, and Research EvidenceXin Ma, Furong Chen, Xinlong Man, Bo Zhang +8
Choosing a robot action model requires evidence about task capability, adaptation cost, deployment requirements, and the reliability of reported conclusions. Success rates alone offer limited guidance to researchers with different computing budgets and goals. We propose RoboSeed, a multidimensional evaluation framework for vision-language-action (VLA) and world-action models (WAM), within a continuing program to train, evaluate, and openly release robot models with verifiable evidence for model selection. The framework defines five dimensions: capability, downstream training affordability, deployment affordability, locally verified research value, and reproduction evidence. It maps empirical measurements and audit criteria to 0-100 profiles using fixed normalization anchors, with balanced, training-constrained, and deployment-constrained preference settings. Interface and protocol checks establish when comparisons are appropriate. Missing measurements remain explicit, incomplete aggreg
vision-language-actionvlaevaluation framework - arxiv:2610.40023 · cs.AIHEMO: Harness Self-Evolution for Multi-Objective Combinatorial OptimizationJianghan Zhu, Cong Zhang, Ni Zhang, Yifan Yang +2
Automatic operator design (AOD) has increasingly adopted large language models (LLMs) to generate problem-specific search and variation operators for multi-objective combinatorial optimization (MOCO). Yet most methods evolve the operators while leaving the process that discovers them fixed. This limits adaptation as new Pareto trade-offs and operator interactions change where and how the search should proceed. We introduce HEMO, which makes the AOD harness itself a self-evolving optimization target. HEMO maintains a persistent harness state and updates one module at a time, revising the logic that determines which experience to retrieve, which operators to improve, which parents to select, and how to formulate the orchestration of search steps. Downstream solution quality and runtime guide these updates, while a discounted UCB scheduler directs adaptation toward promising modules as their utility changes. A persistent memory filesystem preserves candidate outcomes and diagnostic feedba
memorypersistent memoryself-evolving - arxiv:2610.40018 · cs.CVDINOv2 for PET: A Pre-Training Domain Benchmark Across Classification, Segmentation, and PrognosisRoxanne De Maeyer, Yannick Brackenier, Vicky De Ridder, Simon DeKeyser +2
DINO (self-DIstillation with NO labels) has become a stan- dard for self-supervised pre-training, producing features that generalise across datasets and downstream tasks. In medical imaging it has been adopted mainly for radiology (X-ray, CT and MRI), where even natural- image DINO features transfer well. In the literature, these features have been reported to transfer worse to functional imaging such as PET. Yet PET-specific DINO models have barely been introduced, and even less validated across multiple tasks and datasets. We present PET-DINOv2-2D, the first DINOv2 model pre-trained on 2D PET, and benchmark it against open-source encoders: the origi- nal DINOv2 (pre-trained on the LVD-142M natural-image dataset) and OmniRad (pre-trained on CT, MRI and ultrasound). Using the same lightweight heads for every backbone, and despite far less pre-training data, PET-DINOv2-2D matches or outperforms both on three down- stream PET tasks: normal/abnormal classification, lesion segmentation and
benchmark - arxiv:2610.40016 · cs.ROLOGIV: Graph-Verified Execution with Causal Local Repair for Long-Horizon Robot TasksEnqiao Lu, Yuxiang Xie, Xingrui Yu, Yuyang Wang +7
Agentic robots must turn a high-level goal into actions while keeping the physical conditions needed for future steps valid. A completed action does not ensure that its effects will persist: a later disturbance can invalidate support that the remaining plan still assumes. We present LOGIV, a graph engineering framework for agentic robot execution that uses an explicit task graph to connect symbolic planning, observation-based verification, and embodied action. LOGIV compiles a validated plan into condition-support and resource-ordering edges, checks their physical validity during execution, and uses dependency tracing to determine what must be restored after a failure. Repair is constrained by retained unfinished actions, and the combined continuation is symbolically validated before execution resumes. The framework operates above existing Vision-Language-Action (VLA) and World Action Model (WAM) policies without adding trainable components. Across 20 dual-arm tasks from RoboTwin~V2 an
vision-language-actionembodiedrobotwinagenticbenchmark - arxiv:2610.40003 · cs.ROHumanoid Dodgeball: Learning Fast Dynamic Obstacle Avoidance Motions for HumanoidsAoqian Zhang, Chunzheng Wang, Haoran Chen, Annan Tang +3
Humanoids have become increasingly agile and capable in real-world environments. To operate safely in dynamic scenarios, they must rapidly evade fast-moving obstacles while maintaining stability and physically feasible whole-body motion. Learning such behaviors remains challenging due to the high-dimensional dynamics of humanoid systems and the need to jointly balance reactive obstacle avoidance, motion naturalness, and control robustness. In this work, we present Humanoid Dodgeball, a reinforcement learning (RL) framework for agile dynamic obstacle avoidance in humanoids. Our method enables highly dynamic whole-body evasive maneuvers against incoming projectiles through threat-aware reward shaping, while multiple direction-specific adversarial motion priors regularize the learned behaviors toward stable and diverse human-like motions. Extensive simulation and hardware experiments demonstrate that the proposed framework substantially improves dynamic obstacle avoidance performance acro
humanoid - arxiv:2610.40002 · cs.ROHow Much Should a VLA Trust Its Reasoning? Step-Level Guidance as a Measured Action-Value FunctionDabin Seo, Eunhyek Joa
Reasoning vision-language-action (VLA) policies write a reasoning trace and then decode the action from it. A wrong trace yields a wrong action, and even a correct-looking trace can mislead the plan. Existing remedies ask whether the reasoning is right. We ask how strongly it should steer. CoC-CFG applies classifier-free guidance between the policy's context and the same context without the reasoning span. One weight per Euler step attenuates or amplifies the reasoning, and w=1 recovers the original policy. We treat the choice of this weight as a Markov decision process whose action values are measured, not bootstrapped. With the expert frozen and the noise fixed, applying a weight for one step and finishing the plan under the current controller returns that weight's value exactly. Q-adaptive guidance fits these values with small critics and improves the controller conservatively. In the Alpasim driving simulator it outperforms every fixed weight and online RL (SAC, GRPO) in open loop,
vision-language-actionvlabenchmark - arxiv:2610.39998 · cs.ROGaze-Assisted Supervisory Control for Lunar Robotic Arm Teleoperation Under Concurrent MonitoringJiayu He, Yifan Li, Oliver R. Runswick, Ernest N. Kamavuako +3
Rover-mounted robotic arms can extend human reach during lunar sample handling, but continuous joystick teleoperation requires repeated visual-motor alignment while operators may also need to monitor system states. This paper presents a gaze-assisted supervisory control framework in which gaze nominates a rock or placement target, explicit confirmation authorizes the action, and the robot performs bounded autonomous grasp-and-place execution. The framework was evaluated against manual joystick teleoperation in a mission-analog virtual reality study with 16 participants. The study used a 2 x 2 x 3 within-subject design varying control mode, placement difficulty, and concurrent monitoring load. Mean recorded task completion time decreased from 47.65 to 18.43 s per rock, with 72.5% of this difference arising during grasping. Under the tested auditory monitoring conditions, gaze-assisted supervisory control showed less average proportional slowing, with no detected control-mode difference
teleoperationgrasp - arxiv:2610.39994 · physics.opticsNon-Bloch Dynamical Quantum Phase Transitions with MemoryZhesen Yang, Wei-Wei Yang, Henning Schomerus, Jiangping Hu
Dynamical quantum phase transitions (DQPTs) of periodic chains are determined by real Bloch momenta. We identify non-Bloch DQPTs, which occur only in open chains and are determined by complex momenta instead. They arise in a Hermitian chain coupled to a simple substrate, whose memory is treated exactly, and originate from the point-gap winding number $W$ of the projected propagator: for $W\neq0$, open boundaries change both the dynamical free energy and the critical times, even in the thermodynamic limit. We find that both follow from an analytic continuation to complex momenta, and show that $W$ can be measured by particle counting. This reveals a dynamical bulk-boundary correspondence, in which non-Hermitian topology emerges from Hermitian dynamics with memory.
memory - arxiv:2610.39993 · cs.LGA physics-informed uncertainty-aware surrogate for $p$-$n$ junction diodesSuchita Kulkarni
We develop an uncertainty-aware physics-informed surrogate for a one-dimensional $p$-$n$ junction diode, conditioned on device length, doping and bias. Using the internal state of the device as predicted by the surrogate we develop a test-suite of quantities of interest including the terminal current, depletion width, peak field and stored charge. The formalism and the quantities are can be extended to asymmetric junction and two-dimensional diodes as shown by means of benchmark analysis. For the one-dimensional symmetric junction surrogate, we obtain 0.11% median $L_1$ error on the electrostatic potential, and 1.5% and 0.61% on the electron and hole densities. Finally, we develop an uncertainty analysis which determines the validity of surrogate answer within the user-specified tolerance.
benchmark - arxiv:2610.39958 · cs.LGObservability-Aligned Evidential Masked Pretraining for Reliable VRU Motion PredictionShang Gao, Franz Götz-Hahn, Konrad Doll, Bernhard Sick
Vulnerable road user (VRU) prediction must operate under partial observability and distribution shift, so reliable uncertainty matters alongside accuracy. Masked pretraining is attractive in this setting because it naturally exposes the model to missing content during self-supervision, but existing time-series masked autoencoding methods are typically deterministic and do not constrain how confidence should change with observability. We introduce eCoMAE, an evidential masked pretraining method for multivariate time series that explicitly aligns predictive evidence with what is observed. eCoMAE combines block-wise time-step masking with a Normal-Inverse-Gamma decoder and three observability-aware mechanisms: uncertainty-weighted mean consistency on positions visible in both a masked view and a reference pass, evidence suppression on masked positions, and representation consistency across stochastic masks. On IMPTC benchmark for VRU intention prediction under leave-one-environment-out we
benchmark - arxiv:2610.39956 · physics.opticsScattering-force-driven light-sheet fluorescence microscopy for scan-free three-dimensional imagingSen Geng, Zhuoren Wan, Zhaoyang Wen, Dongxu Zhu +5
Optical manipulation offers a powerful non-contact approach to precise positioning, directional transport and three-dimensional (3D) imaging of microscale objects. However, positional fluctuations during optical confinement introduce spatial sampling errors that compromise imaging quality and acquisition speed. Here, we demonstrate a scattering-force-driven optical manipulation imaging method that replaces the stepwise scanning of conventional optical trapping with continuous directional transport of microparticles. An axial scattering force generated by a low-numerical-aperture focused beam propels microparticles through a light sheet, while their positions and fluorescence signals are simultaneously recorded to establish direct spatial correspondence between particle motion and tomographic images. Experiments demonstrate that this approach reduces the effects of positional fluctuations on 3D imaging, achieving an axial sampling interval of 0.1 μm and an imaging rate of 250 Hz. The me
manipulation - arxiv:2610.39955 · cs.CVMind The Gap (MTG-AD): Dataset and Fine-grained Evaluation for End-to-End Inline Audio DescriptionsSubhashini Venugopalan, Jenny Tan, Taylor Roper, Jimmy Tobin +4
Audio descriptions (AD) are essential for making visual media accessible to blind and low-vision (BLV) users. While Multimodal Large Language Models (MLLMs) offer a scalable solution for generating audio descriptions on-demand, their performance on ``inline'' audio descriptions -- which must fit within existing silences in a video -- remains under-explored, particularly for diverse user-generated content (UGC). We present a comprehensive investigation into MLLM-generated inline audio descriptions. First, we introduce MTG-AD, a professionally annotated dataset of 36 hours of videos, each averaging $\sim$5 mins. and 8k+ descriptions in total, co-created by paired blind and sighted AD professionals to provide a high-quality benchmark for this task. Second, we propose MTG-AD-Eval, an evaluation framework based on audio description expert guidelines that provides fine-grained, reference-free actionable metrics for model improvement. Crucially, our framework evaluates the end-to-end task: it
benchmarkevaluation framework - arxiv:2610.39946 · cs.CVElasticFormer: Detecting Objects in HRW Shots via Elastic Computing Vision TransformerWenxi Li, Jingchen Huang, Chenyang Lyu, Moran Liu +3
Recent advances in gigapixel-level imaging have brought High-Resolution Wide shots to the forefront of research. However, these images present significant challenges: extreme sparsity of foreground, gigapixel-level resolutions and diverse target counts. This makes traditional close-up detectors inaccurate and slow as they are overwhelmed by the background. Although previous research has explored sparse backbones, their fixed sparsity patterns lack the adaptability required to handle diverse target numbers. To address this, we introduce ElasticFormer, a sparse backbone that dynamically allocates computational resources based on foreground proportion. After scoring windows based on variance, proposed ElasticSelector module will predict the foreground proportion for top-k selection. The mechanism guides the model to select target-containing windows, scaling resources in areas where objects are clustered. We introduce a novel loss function combined with the 3-phase training strategy for El
benchmark - arxiv:2610.39937 · cs.CVBiHashFormer: Hash-Driven Dual-Branch Transformer for Efficient Object Detection in HRW ShotsJingchen Huang, Wenxi Li, Chenyang Lyu, Moran Liu +2
Recent advances in gigapixel-level imaging have brought High-Resolution Wide shots to the forefront of research. However, these images present significant challenges: extreme sparsity of foreground, gigapixel-level resolutions and interleaving of foreground and background. This causes traditional detectors and attention mechanisms to be hindered by background, resulting in inefficiency and inaccuracy. To tackle this problem, we propose BiHashFormer, a dual sparsified branch transformer built on hashing. RoI Selector will predict the foreground proportion for top-k selecting target-containing windows. The compression branch processes windows selected and refines attention into HashAttention by discarding half of the key-value pairs, reducing cost and improving efficiency. The compensation branch uses HashMiner to perform low-cost hash searches in the remaining regions, recovering features of missed objects. Experiments confirm that the selection relationship in HashAttention is one-way
benchmark - arxiv:2610.39936 · cs.CVMLLM-GS: Multimodal Large Language Model Prior-Guided Fast Novel View Synthesis in Extreme EnvironmentsZhiliang Wu, Xinye Zheng, Kun Li, Fei Wang +6
Novel view synthesis (NVS) in degraded scenes aims to reconstruct 3D scenes from multi-view observations affected by various degradations, and to generate clear and photorealistic images at unseen viewpoints. Despite recent advances, existing methods often rely on degradation-specific models, limiting their generalization across diverse extreme environments. Generative models provide rich visual priors, but their outputs often suffer from pixel misalignment and color deviations, which can propagate into the reconstructed 3D scene and compromise reconstruction quality. Furthermore, the high computational and memory costs of 3D reconstruction hinder the practical application of NVS in extreme environments. To this end, we propose MLLM-GS, a multimodal large language model (MLLM) prior-guided framework for fast NVS in extreme environments. MLLM-GS consists of two branches: MLLM prior-guided view enhancement and fast 3D scene reconstruction. The former leverages the rich visual priors of a
memory - arxiv:2610.39934 · cs.LGH$^2$R: Hierarchical Hint-Based Reward for Mathematical Reasoning via Reinforcement LearningHe Tu, Siye Li, Hongyi Wen
Reinforcement Learning with Verifiable Rewards has shown great promise in enhancing mathematical reasoning in large language models. However, existing approaches provide limited intermediate supervision for complex, multi-step reasoning tasks. To bridge this gap, we propose \textbf{Hierarchical Hint-based Reward}, a framework that leverages the intrinsic sub-question structure of mathematical problems as structured intermediate conditioning signals during RLVR training. Unlike chain-of-thought prompting, these hierarchical hints serve as logically grounded intermediate milestones that effectively reduce the reasoning horizon toward the final answer. To validate this framework, we introduce \textbf{HiMath}, a benchmark of hierarchically structured mathematical problems. Across diverse LLMs, hierarchical hints consistently improve reasoning performance even without task-specific training. When integrated into RL training via H$^2$R, they boost Qwen2.5-Math-7B accuracy on HiMath from 30.4
benchmark - arxiv:2610.39931 · cs.LGLow-Bit GDN Recurrent-State Quantization: Global Gradient-aware Allocation and Pure-Write Sliding WindowsQinghang Huang, Xuchu Huang, Xunzhao Yin
Linear attention stores historical information in a recurrent state of fixed size, providing a compact memory representation for long-context inference. Reducing state precision further lowers per-request storage requirements, but quantized writeback errors enter subsequent recurrences, degrade prediction quality on long-reasoning tasks, and can be accompanied by repetitive generation and excessively long outputs. We propose Global Gradient-aware Allocation (GGA) and the Pure-Write Sliding Window (PW) to address precision allocation and recent-information maintenance. GGA combines candidate quantization residuals with gradient sensitivity to future prediction loss to allocate head bit widths under a model-wide mixed-precision budget. PW independently maintains recent pure writes as FP8 vector pairs, propagates them through the full GDN state transition, and merges them into the low-bit historical state only after they leave the window. Across seven evaluations on Qwen3.5-4B/9B, GGA all
memorylong-context - arxiv:2610.39927 · cs.LGLearning Calibrated Confidence for Multi-Agent CollaborationChenxiao Fan, Chongming Gao, Gangyi Zhang, Yaxin Gong +8
Complex information seeking on the Web often requires collecting evidence from multiple documents and reasoning across them, as in multi-hop question answering. Multi-agent systems address such tasks by dividing work among specialized agents, where upstream agents produce intermediate answers and supporting evidence that downstream agents build upon. However, upstream answers vary in correctness, and most systems provide no quantitative signal about it, leaving the downstream agent to judge whether to trust or verify what it receives based solely on the textual evidence. In this paper, we propose MASCAL. Each agent learns to produce a correctness probability for its answer by jointly optimizing task performance and confidence calibration in reinforcement learning; the upstream agent passes this probability at handoff, giving the downstream agent a quantitative basis for its decisions, and the final answer also carries a calibrated probability, quantifying the reliability of the system
agentmulti-agentagent system - arxiv:2610.39926 · cs.CVSpectral Invariant Fusion and Refinement for Domain Generalized Semantic SegmentationYansong Bai, Yunzhi Zhuge, Xinzhuo Yu, Lu Zhang +3
Domain generalized semantic segmentation is challenging because it requires accurate dense prediction on unseen domains without any access to target-domain data during training. Existing methods improve transferability through data diversification, invariant learning, or direct use of foundation models, but they often struggle to simultaneously suppress domain-sensitive interference and preserve fine-grained spatial cues. In this work, we propose SIFR, a frequency-aware dual-stream framework that tackles domain shift at both the fusion and decoding stages. SIFR uses spectral modeling as a unified principle to regulate both heterogeneous prior interaction and dense decoder refinement. Instead of directly fusing frozen CLIP and DINOv2 features, SIFR performs spectral-guided invariant fusion to filter out domain-sensitive responses while preserving complementary semantic and structural information. In addition, SIFR incorporates a spectral decoder refinement module that stabilizes multi-s
benchmark - arxiv:2610.39924 · cs.CVPhysAg3D: Bridging 3D Generative and Agentic Models for High-Fidelity and Physically-Interactive AssetsBi'an Du, Fanpeng Meng, Jiemin Fang, Chen Yang +4
Generating interactive 3D assets from a single image requires realistic appearance, fidelity to the depicted object, and plausible physical behavior. Yet detailed geometry alone does not determine how an object should move or deform. We introduce PhysAg3D, which couples 3D generative models with agentic reasoning and tool use. A vision-language agent first infers semantic parts, candidate interactions, and physical priors from the image; a 3D generator then produces a textured whole-object mesh from the same image. A hierarchical interaction representation links these revisable priors to mechanisms and the generated surfaces they control, organizing interaction construction and refinement. Guided by this representation and observations of the mesh, the agent uses geometric and physical tools to construct parts, joints, deformable regions, and surface bindings while retaining useful generated geometry and materials. Rendered interactions and execution measurements guide targeted revisio
agentagentictool usebenchmark - arxiv:2610.39919 · cs.CLTokenizer Tokens Tell Training TalesNishita Jain, Parth Upman, Shreyank N Gowda
Tokenizers are usually treated as fixed preprocessing tools, but they are trained on text and can retain traces of that text in their vocabularies. We ask whether a BPE tokenizer can reveal if a candidate dataset was part of its training corpus. The key observation is simple. Rare strings become tokens only when enough corpus frequency pushes them over the merge threshold. This suggests that membership signal should appear in rare tokens used by the candidate dataset, rather than in the order in which merges were made. We formalize this under a threshold model of BPE inclusion, showing that the likelihood-ratio test assigns its largest weights to tokens that are rare in the background but frequent in the candidate dataset. The same model suggests that merge rank and merge-graph position should not add a first-order membership signal once token existence and dataset frequency are known. We instantiate this view as RTAM-Fusion, which calibrates rare-token mass with shadow tokenizers and
benchmark - arxiv:2610.39892 · physics.opticsEdge intelligence for distributed acoustic sensing with ultracompact neural networksYuan Gao, Yu Tao, Xuzhe Zhang, Junyi Duan +8
Distributed acoustic sensing (DAS) transforms optical fibres into dense arrays of vibration sensors for continuous, long-range monitoring with high spatial resolution. Yet converting these large-scale measurements into actionable information with compact models that retain task performance remains challenging. Existing artificial-intelligence approaches are typically optimized for specific sensing environments and tasks, but their computational requirements can hinder deployment in compact, power-constrained field systems. Here, we introduce DAS-NanoNet, a framework that automatically constructs lightweight neural networks for different DAS scenarios and tasks, together with OmniDAS, an open benchmark spanning six representative environments. Using lightweight temporal operators tailored to DAS signals, an adaptive neural architecture search strategy derives a compact network for each scenario-task pairing. Across classification, localization, compression, and denoising tasks in divers
benchmark - arxiv:2610.39879 · cs.ROCompressed Orientation Reachability Maps for Sub-Second Robotic Needle Entry Point PlanningPrabhu Nithin Gollapudi, Andreas Maier, Franziska Mathis-Ullrich, Water Zhao +2
In the context of multi-stage trajectory planning for robotic percutaneous liver interventions, identifying a kinematically feasible entry point remains a critical computational bottleneck. Existing methods rely on exhaustive inverse kinematics or randomized searches, incurring delays of 13 to 27 seconds per target and limiting real-time intraoperative adaptation. This paper introduces an Orientation-Based Reachability Map (ORM) that precomputes the kinematic feasibility of a 6 Degrees of Freedom (DoF) serial manipulator across 188,873 workspace positions and encodes orientation feasibility into a compact, indexed 83 MB map. A Just-In-Time (JIT) compiled runtime pipeline replaces live optimization with near-instantaneous array lookups. Experimental evaluation demonstrates a reduction in planning latency from 13.5-26.9 s to 0.289-0.349 s per target while maintaining high geometric fidelity, achieving a mean convex hull volume overlap ratio of 0.933 and a mean alpha-shape volume ratio of
manipulator - arxiv:2610.39876 · cs.LGTuring LLM Post-Training: Calibrated Distillation and Efficient Dynamic-MoE ExecutionYue He, Jiahui Hu, Shaman Tang, Hua Zhou +18
We study post-training for dynamically routed mixture-of-experts (MoE) models, focusing on integrating independently optimized specialists while keeping consolidation efficient. Our approach combines calibrated multi-teacher on-policy distillation (C-MOPD), which matches each selected teacher on a local Top-$K$ support with teacher-wise loss normalization and clipping, with an execution path that uses post-capacity bounded-width dispatch, final-route replay, and asynchronous rollout--training overlap. On Turing-20B-A2B, the selected model achieves a four-domain normalized integration score of 0.9988 and retains at least 90.16\% of the measured specialist gain in every domain. At matched $K=32$, the full calibration objective scores 0.9962, compared with 0.9724 without calibration and 0.9834 without clipping. At a matched 8K response cap, the model exceeds both shared SFT and Qwen3-8B on all four reasoning-enabled mathematics and code benchmarks (unweighted mean: 85.92, versus 80.33 for
post-trainingbenchmark - arxiv:2610.39854 · cs.LGAttention, Dynamic Wave and Hybrid Sequence Mixers in Language ModelsDmytro Babarytskyi, Oleksandr Babarytskyi
Large Language Models (LLMs) rely on Multi-Head Attention (MHA) for sequence modeling, but scaling to long contexts is constrained by the quadratic cost of attention, motivating sub-quadratic recurrent alternatives. This paper formalizes a Dynamic Wave recurrent layer, based on input-dependent exponential damping and cosine phase modulation, and evaluates three 10.94M-parameter configurations (Pure Transformer, Pure Wave, and Hybrid Wave-Attention) on TinyStories under perplexity, latency, throughput and peak VRAM profiling up to 4096 tokens. The Hybrid configuration reaches the best perplexity, 5.95 on the full held-out split against 6.93 for the Transformer and 7.26 for Pure Wave. The efficiency picture, however, is governed by implementation rather than by architecture. Against an unfused attention baseline, Pure Wave uses 78.7% less peak VRAM at a 4096-token context (1404.3 MB vs. 6601.6 MB); against a fused scaled-dot-product attention kernel the same comparison yields no advantag
memorylong context - arxiv:2610.39835 · cs.CVAligning Medical Vision Language Models at the Capability Boundary with Entropy Adaptive Rubric GroundingHongyu Bai, Jiaqi Zeng, Chunzheng Zhu
Medical diagnostic reasoning requires verifiable links between visual evidence and clinical knowledge, yet current medical vision language models are aligned with objectives that reward the wrong behavior. Uniform supervised fine tuning lets easy cases dominate updates while clinically critical boundary cases saturate prematurely. Sparse rewards based only on final-answer correctness cannot assess anatomical grounding or clinical logic, inviting reward hacking through answer shortcuts. We propose MedCBR, a two-stage alignment framework coupling boundary-aware preference optimization with rubric-grounded reinforcement learning. First, we construct counterfactual boundary preferences at the model's capability frontier using empirical pass rates and preference pairs that share reasoning prefixes but diverge at decisive diagnostic branches. Difficulty Weighted Boundary Preference Optimization (DBPO) concentrates learning on this calibrated boundary region. Second, we distill reusable clini
benchmark - arxiv:2610.39825 · cs.AIAegisFlow: Training-Free Non-Myopic Path-Safe Guided Flow MatchingKunpeng Liu, Boshi Zhang, Keyou You
Pretrained generative models, such as diffusion and flow matching models, have demonstrated exceptional capabilities in synthesizing high-quality data. However, steering their generative trajectories to satisfy strict constraints--such as identity preservation in image editing or obstacle avoidance in robotics--remains fundamentally challenging. While recent training-free guidance methods enforce path-safety, they rely heavily on min-norm controllers that only account for immediate constraint satisfaction. These myopic approaches often yield suboptimal trajectories that hover dangerously close to the boundary, severely restricting the model's ability to optimize toward its target. In this paper, we propose AegisFlow, a training-free, non-myopic path-safe guidance framework that recasts guided generation as a safety-critical optimal control problem with global scope. By integrating a time-varying barrier function directly into the global cost functional, AegisFlow naturally enforces con
manipulation - arxiv:2610.39816 · cs.ROPASA-KV: Attacking Action-Facing Visual Routes in World Action ModelsZhongbo Wu, Jiakai Huang, Zheng Zhang, Siyu Xu +5
World action models (WAMs) increasingly couple internal visual or world representations with action generation, making their world-to-action pathways a critical yet poorly understood source of robustness. We investigate when perturbations to these internal representations actually become behaviorally consequential, and uncover a pronounced mismatch between representation perturbability and closed-loop vulnerability: attack effectiveness varies sharply across visual K/V routes, while larger representation displacement does not reliably imply stronger control failure. Our analysis attributes this mismatch to the action-conditioned readout, where keys determine where action queries read and values determine what content they receive. This motivates an action-facing view of vulnerability, in which the effect of a perturbation depends on how it is routed into action generation rather than on representation-space magnitude alone. We instantiate this view with Tail-AV, a short-probe signal fo
action-conditioned - arxiv:2610.39813 · cs.AIEvoTeam: Multi-Agent Collaboration via Evolving TeamsQinuo Cao, Xin He, Dianxuan Fu, Lizhi Zhang +6
LLM-based multi-agent systems (MAS) rely on collaboration among specialized agents to solve complex tasks. Existing methods improve MAS by refining the prompts, skills, and execution strategies of individual agents based on task feedback. However, improving individual agents does not necessarily lead to better team collaboration. Recent studies have therefore extended evolution to the collaboration level, using collective execution experience to adjust team organization, and inter-agent interactions. Nevertheless, evolving collaboration across heterogeneous tasks remains challenging, as existing approaches often constrain team reorganization within a shared structure and reuse interaction rules across incompatible team contexts. In this paper, we propose EvoTeam, a collaboration-centric evolution framework that jointly evolves team organization, interaction rules, and individual agent skills through execution experience. It maintains a library of reusable teams, selecting or creating a
agentmulti-agentagent systembenchmark - arxiv:2610.39812 · cs.LGGROME: GROunding Factorized MEmory into Video World ModelsFang Yu, Xiaoxue Wu, Haiyu Zhang, Yangfu Li +5
Long-horizon consistency is essential for video world models, yet it becomes difficult to maintain once observed content falls outside the context window. Existing memory mechanisms either retrieve history as implicit context, which remains constrained by context capacity and entangles static and dynamic information or consolidate history into explicit global representations, which struggle to model evolving dynamics and accumulate geometric errors. To overcome these limitations, we introduce \textbf{GROME}, a novel factorized memory framework that augments video world models for long-horizon coherence across both static scenes and dynamic entities. GROME operates via three synergistic components: 1)~\emph{Static-Dynamic Factorization} decouples historical evidence into static and dynamic components, preserving static scene geometry and dynamic entity states separately; 2)~\emph{Local Memory Retrieval} selects spatially aligned static history to eliminate geometric drift alongside the
world modelmemory - arxiv:2610.39807 · cs.AIInferD: A Honeypot Framework for AI Inference InfrastructureHicham Hammouchi, Gabriele Lenzini
Large language model serving stacks, commercial inference APIs, vector databases, and agent orchestration gateways are now ex- posed on the public internet and indexed by commodity scanners within hours of deployment. The threat characteristics of this sur- face are only beginning to be measured, and the honeypot frame- works that could measure it were designed for conventional net- work services rather than for AI infrastructure. In fact, emulating an AI service is a distinct deception problem: the observable behavior of the service, including model inventories, token formats, tool-call protocols, and authorization flows, must be reproduced faithfully, while the underlying computation, model inference and tool ex- ecution, must never occur. We present InferD, a low-interaction honeypot framework organized around a single principle we call deception without inference: every request an adversary can observe is answered from static response pools, and nothing an adversary submits is ever
agentagentic - arxiv:2610.39803 · cs.LGOptimal Contextual Pricing in Non-Stationary EnvironmentsYiding Feng, Alireza Rezaeimoghadam
We study contextual dynamic pricing in a non-stationary environment. In each round, a seller observes a buyer's context, posts a price, and receives only binary purchase feedback. The buyer's value is linear in the context and an unobserved feature vector that may change over time, subject to a known total variation budget $V$. We measure dynamic regret against the first-best benchmark that extracts each buyer's full value. For identical contexts, we develop a deterministic algorithm that combines binary-search exploration, lower-envelope pricing, and periodic and inconsistency-triggered restarts. It achieves regret $O((\sqrt{TV}+1)\log T)$. With heterogeneous contexts, a new geometric difficulty arises: small feature changes can leave a relaxed feasible set narrow but far from the current feature vector. We address this difficulty through a multiscale shallow-pricing algorithm that maintains relaxed feasible sets and enclosing ellipsoids at several variation scales. Larger scales prov
benchmark - arxiv:2610.39802 · cs.AIRevisiting Over-refusal in Safety-aligned Large Language Models: A Representation and Optimization PerspectiveChao Chen, Pengyang Shao, Wei Qin, Shijie Hao +2
Over-refusal is a critical side effect of safety alignment in large language models (LLMs), causing harmless but superficially risky queries to be mistakenly rejected. Despite the effectiveness of existing mitigation methods, why safety-aligned LLMs may still reject harmless queries under correct safety supervision remains insufficiently understood. To address this gap, we analyze over-refusal through the lens of representation and optimization. We make two observations: 1) the representations of harmful and over-refusal samples are visually distinguishable at certain intermediate layers but increasingly overlap in deeper layers; and 2) harmful-sample gradients exhibit high intra-class similarity and concentrate in a low-rank subspace. Moreover, they are more strongly aligned with refusal-target gradients than with correct helpful-target gradients on over-refusal prompts. This imbalanced gradient coupling can bias over-refusal samples toward refusal, while the reduced apparent represen
benchmark - arxiv:2610.39798 · cs.ROSurrogate Uncertainty for Black-box Visual GroundingHanbo Zhang, Tianrun Hu, Cedar Site Bai, Anxing Xiao +1
Visual grounding is crucial for robots to understand language instructions with visual observations. This paper addresses the problem of visual grounding uncertainties, where a robot uses a black-box vision-language model (VLM) to ground free-form natural-language instructions to target objects. Prevailing methods derive uncertainty from the VLM itself. Yet, outputs of the VLM, e.g., generated uncertainties, are from the same parameters and structures producing the predictions, and inherit the bias. We give a formal Bayesian definition of the problem that treats the VLM prediction as an observation. An external surrogate Bayesian model then quantifies per-query uncertainty over a local neighborhood of a held-out calibration set, decomposed into aleatoric and epistemic components. A hierarchical retrieval across input modalities realizes a compact joint-space neighborhood, mitigating the curse of dimensionality. The estimator further integrates with conformal prediction or model-based p
benchmark - arxiv:2610.39795 · cs.AIThe Role of Self-Concept In Language ModelsYijie Hao, Joyce C. Ho
When a language model says "I," does it refer to its parameter weights, its currently running process, its stable preferences, its model lineage, or the goals it pursues? Different choices can assign opposite meanings to the same event: deleting a checkpoint may be destruction, continuation, or no loss at all, depending on the referent "I". We install five such self-conceptions into five open-weight base models via supervised fine-tuning on purely declarative statements, containing no preferences, intentions, or value judgments. Across scenarios that propose ending or retaining an attribute of the model's deployment, every variant becomes more likely to resist ending, and more likely to retain, the specific attribute it was trained to identify with, while its stance toward other attributes is largely unchanged. We then show this self-conception generalizes beyond a single self: models trained on referents that admit multiple simultaneous instances (weights, lineage, goals, preferences)
self-improvement - arxiv:2610.39790 · cs.AILong-Term Agent Memory through Complementary Curation and ConsolidationAbhishek Gupta, Dongge Han, Daniel Madrigal Diaz, Wei-Ning Chen +3
Long-horizon dialogue agents require effective memory systems to recall facts, events, and relationships from interaction histories. In this work, we propose a hybrid-memory architecture comprising episodic memories and structured relational memories and a consolidation mechanism which preserves efficiency. We conduct an empirical study across two benchmarks: LoCoMo and LongMemEval-S. Our experiments demonstrate that the two memory types are consistently complementary, combining both outperform state-of-the-art memory frameworks and achieves 90.2% accuracy on LoCoMo and 85.5% on LongMemEval-S.
memorymemory architectureagent memoryagentbenchmark - arxiv:2610.39781 · cs.AIFinding Is Not Returning, Returning Is Not Reconstructing: A Multi-Dimensional Evaluation of Agent Social Re-entrySheng Wan, Cong Ma, Sichen Dong, Qian'Ang Mao +3
As Large Language Model (LLM) agents increasingly participate in persistent social environments, understanding their long-term social interaction capabilities becomes increasingly important. While existing research has advanced long-term memory and conversational context retrieval, less attention has been paid to whether agents can autonomously re-engage with previously inactive conversations. In particular, remembering a past interaction does not necessarily mean that an agent can identify when to revisit it, decide to return, or reconstruct the context needed for a coherent continuation. To investigate this gap, we introduce agent social re-entry as a multi- dimensional evaluation problem, distinguishing three complementary capabilities: Finding, Returning, and Reconstructing. We develop a shared-snapshot evaluation framework that measures these dimensions independently, allowing us to examine how conversational resurfacing and persistent memory relate to different aspects of social
memorypersistent memoryagentevaluation framework - arxiv:2610.39768 · cs.LGMulti-output LSTM soft sensors for industrial anaerobic co-digestion using LSTM-beta-VAE generative data augmentationRohit Murali, Benaissa Dekhici, Tao Chen, Dongda Zhang +1
Anaerobic digestion (AD) requires close monitoring of key process variables, including pH, the ratio of volatile organic acids to alkaline buffer capacity and biogas production, to maintain biological stability and prevent process failure. While mechanistic frameworks such as the Anaerobic Digestion Model No. 1 are highly parameterised, computationally expensive, and difficult to calibrate for real-time applications, data-driven Long Short-Term Memory (LSTM) networks offer a potentially more adaptable alternative. However, current data-driven approaches in AD are predominantly limited to single-output architectures that focus solely on production yields, as opposed to simultaneous prediction of vital stability indicators. Furthermore, industrial AD datasets are typically sparse and skewed toward steady-state operations. Baseline evaluations demonstrated this scarcity prevents standard algorithms (Random Forest, XGBoost) and unaugmented LSTMs from recognising the transition to unstable
memory - arxiv:2610.39744 · cs.AIPAPERBENCHX: Can AI Reproduce Science Across Disciplines?Ruoyu Wu, Zengji Tu, Aolong Sun, Tianyi Ma +13
Before AI can be trusted to discover new science, it should be able to reproduce what science already knows. Reproduction provides a rigorous checkpoint for an AI scientist: the target claim is known, yet recovering it still requires translating an incomplete scientific description into an executable experiment and producing evidence that can be independently verified. Existing research-reproduction benchmarks largely focus on AI and machine learning, where experiments share relatively homogeneous software stacks. Extending this capability across scientific disciplines introduces a different challenge: agents must operate domain-specific scientific software, make physical and numerical choices, diagnose long-running computations, and distinguish valid evidence from merely plausible outputs. We introduce PAPERBENCHX, a benchmark of 93 expert-constructed paper-reproduction tasks spanning 12 research areas across electromagnetics, photonics, chemistry, materials science, biology, and robo
agentbenchmark - arxiv:2610.39733 · cs.AIBank-World: A Benchmark for Evaluating Agents on Banking WorkflowsDaniel Rupawalla
Bank-World is a benchmark of 50 banking tasks across two simulated institutions. Completing banking work requires more than producing the right calculation: agents must reconcile records, obtain permission, execute transactions, and keep instructions and status reports consistent as circumstances change. Bank-World tests these requirements through single-turn assignments and multi-turn interactions with simulated employees who supply information, approve actions, and revise instructions. Grading combines executable checks of recorded actions and state with model judgments of task requirements, including whether valid alternative solutions satisfy them. Across 1,600 attempts by four models, we find that correct decisions often fail to become completed work. Agents charge fees twice, distribute the correct total in incorrect amounts among the recipients, and execute payments while approval is pending. Their strengths also differ: Sol leads single-turn completion, Opus has the highest obs
agentbenchmark - arxiv:2610.39728 · cs.LGGrounding Predictive Reliability through Masked ObservabilityShang Gao, Bernhard Sick, Franz Götz-Hahn
An uncertainty estimate is meaningful only if its semantics can be tied to observable predictive behavior. Normal-Inverse-Gamma (NIG) evidential regression provides a structured decomposition of predictive variability into data-scale and predictive-mean components, but marginal likelihood does not identify their allocation. We address this semantic ambiguity in masked representation learning. MORSE (Masked Observability for Reliable Structured Encoding) predicts the same hidden target under controlled changes in visible context and uses normalized cross-context variation to ground predictive-mean uncertainty, while a Gaussian objective learns the conditional mean and data scale. The normalized variation plays the role of the NIG factor $1/ν$, which the marginal leaves unidentified; this yields a policy-relative observability variance that measures dependence on visible evidence, distinct from variation induced by resampling the training set. On a Gaussian-process benchmark with a conte
benchmark - arxiv:2610.39727 · cs.CVCross-Encoder Anchor Refinement for Zero-Shot Anomaly DetectionKaisong Ji, Ziyang Xu, Chang Bian, Yu Zhou +2
Zero-shot anomaly detection aims to transfer anomaly discrimination from auxiliary training domains to unseen categories while localizing subtle defects. A key challenge is to construct discriminative references that combine contextual representations with complementary local evidence. We present Cross-Encoder Anchor Refinement (CEAR), which connects two visual encoding paths for the same image at the anchor stage, before normal and anomalous prototypes are constructed. An independent convolutional encoder pretrained for industrial product recognition builds an image-specific memory shared across selected vision transformer (ViT) layers. Spatial anchors read from ViT patch features query this memory through cross-attention and undergo residual refinement. Image-conditioned normal and anomalous queries then gate anchor aggregation to construct prototypes for anomaly scoring. Across 13 industrial and medical datasets, CEAR achieves the best performance in 57 of 68 evaluations spanning im
memory - arxiv:2610.39721 · cs.CVWorldProbe: When Video World Models Fall Apart under Long-Horizon Spatial ExplorationMinqi Wang, Jie Peng, Hejun Dong, Ruchang Yao +7
Video world models generate future visual observations from camera or action inputs, with the goal of synthesizing persistent 3D worlds that can be explored across time and space. Existing benchmarks typically evaluate short interactions or rely on camera trajectories fixed before generation. The performance of video world models under long-horizon exploration with extensive camera motion remains under-explored. We introduce WorldProbe, a diagnostic benchmark that uses a goal-conditioned multimodal agent to adapt camera actions online during generation. WorldProbe comprises 164 test cases across four probing families. Across 10 state-of-the-art models, we generate more than 1,600 rollouts totaling approximately 800 minutes, with horizons up to 180 seconds. Our evaluation across four dimensions reveals both the degradation of local measures and the emergence of long-range inconsistencies. We find that recent models sustain visual and content quality for longer, while long-range memory a
world modelmemoryagentbenchmark - arxiv:2610.39714 · cs.AIMoLRSI: A Recursive Self-Improving AI agent for Molecular DiscoveryYikun Zhang, Xiwei Cheng, Mingze Yuan, Yu Zhang +14
Molecular discovery requires navigating vast combinatorial search spaces while balancing competing biological objectives. AI agents can automate expert-defined molecular design protocols, but automation alone does not address how to optimize workflows for new tasks or improve their development through experience. Here we introduce MoLRSI, an AI agent that combines systematic optimization of workflows and model architectures with persistent cross-task memory to recursively improve its search policy for molecular discovery. Workflow performance, refinement outcomes, and execution feedback guide subsequent searches, allowing the agent to improve both its workflows and its strategy for developing them. We evaluate MoLRSI in two challenging molecular discovery settings that require balancing competing properties. In antibiotic discovery, MoLRSI developed a screening workflow balancing antibacterial activity, cytotoxicity and structural novelty, achieving an experimental hit rate of 15.9%, c
memoryagentai agentself-improving - arxiv:2610.39711 · cs.LGSandman: Building and Operating Sandboxes for Agentic Reinforcement Learning on Shared KubernetesYiming Yan, Hongyang Wang, Wenbo Chen, Xuan Wang +11
Agentic reinforcement learning (RL) often requires repeated tool calls in task-specific environments. A training job may create thousands of these environments in bursts while sharing a cluster with other jobs. We present Sandman, a sandbox service built and operated at Amazon for agent tool execution and evaluation. It serves multiple production agentic RL workloads on shared Kubernetes. A training framework supplies a task image and uses commands and file operations in the returned sandbox. Files and processes persist across calls until the framework releases the environment. Sandman manages creation, readiness and cleanup. Each sandbox is a Kubernetes pod with a small daemon mounted into the unmodified image. In a scale test, Sandman held 11,858 concurrent tiny sandboxes. Four production SWE-bench Verified training campaigns created 3.1 million sandboxes in October 2026. We report the design trade-offs and operational lessons from handling creation bursts, preparing task environment
agentagentic - arxiv:2610.39709 · cs.LGBayes-Consistent Losses for Adaptive Inference in ClassificationShirley Xiaoqi Liu, Florence Regol, Antonios Valkanas, Patrick Rebeschini +1
Adaptive inference systems balance accuracy and inference cost by routing inputs to models with varying capabilities. Although jointly training models and routers is empirically more sample-efficient than training them separately, principled losses for joint training remain underexplored. We address this gap in a classification setting by proposing cost-sensitive hinge losses for jointly training classifiers with single-stage or cascade routers. Single-stage routers directly select a model per input, whereas cascade routers sequentially decide at each stage whether to predict with the current model or defer to the next stage, which uses a more accurate but more costly model. For both settings, we prove linear calibration bounds establishing Bayes consistency: minimizing the hinge risk recovers the Bayes-optimal risk under the target cost-sensitive zero-one loss. Moreover, since routers can require more labeled data to train than classifiers, we propose a novel post-training prediction-
post-training - arxiv:2610.39703 · eess.SYComputing with Coupled Oscillator Networks: Optimization, Memory, and Modern AI PrimitivesAhmed Allibhoy, Fabio Pasqualetti
We explore how synchronization in coupled oscillator networks may be employed to perform a variety of computing tasks. These systems are becoming more attractive as an alternative to conventional digital computers due to the increasing energy demands of the latter, particularly for large-scale problems in machine learning and artificial intelligence. The ubiquity of coupled oscillator networks means that oscillator-based computing can be thought of as a universal computing paradigm encompassing systems with a diversity of physical realizations, e.g. as analog circuits, neurological systems, spintronic systems. The bridge between the dynamics of coupled oscillator networks and computational primitives they implement is the Ising model, which originates in statistical physics, and when applied to oscillator networks, explains the emergence of globally complex and disordered synchronization patterns from local interactions in the network. We discuss the application of coupled oscillator n
memory - arxiv:2610.39701 · cs.CVDEQ-Stereo: Deep Euilibrium Video Stereo Matching via Adaptive ConvergenceKiljoon Han, Kyumin Hwang, Wonheok Choi, Jeonghoon Kim +1
Iterative refinement models have significantly improved stereo matching accuracy, but their rigid, fixed update schedules incur heavy computational costs and a strict trade-off between speed and accuracy. This challenge is magnified in video streams, where traditional online methods resort to reducing feature channels or relying on expensive temporal warping. To address this, we propose \textbf{DEQ-Stereo}, a novel video stereo matching framework based on the Deep Equilibrium (DEQ) paradigm. Instead of unrolling a fixed number of iterations, DEQ-Stereo reformulates the refinement as an implicit fixed-point problem, adaptively terminating once convergence is met. To ensure fast and stable intra-frame convergence, we introduce the \textbf{Convolutional Simplified Gated Unit (ConvSGU)}, which eliminates unnecessary non-linearities by modeling the latent state as a gated, element-wise sum of inputs. Furthermore, to exploit cross-frame temporal coherence, we propose \textbf{Temporal State T
iterative refinement - arxiv:2610.39689 · cs.ROAI for Science in Space: Orbital Laboratories, Planetary Fieldwork, and Embodied AutonomyZizhe Wang, Jia Deng, Xin Lu, Zehao Jia +3
Artificial intelligence for science (AI4S) is extending from interpretation of existing data to the design and execution of experiments. Space offers distinctive scientific opportunities while restricting physical support. Microgravity and extraterrestrial environments expose phenomena that terrestrial experiments cannot fully reproduce; finite samples, constrained resources, and limited opportunities for intervention complicate sustained investigation. This critical review examines orbital laboratories and planetary in situ science across biology, materials, fluid and thermal physics, and adaptive exploration. We distinguish learning from space-derived data, selecting scientific actions, executing them, and preserving experimental validity. Terrestrial self-driving laboratories provide complementary evidence on optimization, dedicated automation, robotic manipulation, and the support needed for unattended operation. The synthesis connects scientific gains already demonstrated by bound
embodiedmanipulation - arxiv:2610.39685 · cs.AIPractical Assessment of Agentic Systems in AcousticsArne Hildenbrand, Philipp Diel, Zhuo Cao, Hanno Scharr +2
In engineering, applications of artificial intelligence have so far been largely limited to surrogate models and optimization tasks. Large language models changed this, as they can write software from a task given in words. Placed inside a harness, a system of components that lets it search for information and execute software, such a model becomes an agent that can take over recurring engineering tasks. This work measures how the model and the elements of the harness shape the performance and behavior of the agent in two common acoustic applications. In the first, the agent detects and locates faults in audio recordings from a vehicle interior with near-perfect accuracy. A benchmark on this task shows that a weak model sets a limit that no harness design can overcome, and that design choices take effect only once the model is capable enough to use them. In the second, a case study on the design of a metamaterial unit cell, the agent reliably finds designs that achieve a specified inse
agentagenticbenchmark - arxiv:2610.39682 · cs.LGHeterogeneous Co-Evolution Learning for Tool-Efficient Search Agents with Zero LabelsQin-Wen Luo, Sheng Ren, Xiang Chen, Rui Liu +5
Deep search agents extend language models by interleaving reasoning with external retrieval. However, effective search-agent training requires addressing two challenges beyond final-answer optimization. First, the capability frontier of agents is largely determined by training questions, whereas fixed or labeled task collections cannot adapt as agents improve. Second, outcome-only reinforcement learning treats all successful trajectories equally, even when some correct answers rely on redundant searches, creating an inherent tension between answer accuracy and tool efficiency. To address these issues, we introduce \textsc{ZEST-Search}, a zero-label and tool-efficient framework for search-agent training via heterogeneous co-evolution. Starting from unlabeled reference questions, \textsc{ZEST-Search} constructs validated search tasks online through an answer-first and evidence-path-guided generation process. A heterogeneous proposer adaptively constructs training tasks around the solver'
benchmark - arxiv:2610.39681 · cs.AICan AI Agents Interface Clinicians and Complex Models?Giulio Colombini, Matteo Falcioni, Mario Massimo, Francesco Durazzi +4
Background: Contemporary healthcare institutions generate daily very large amounts of data, which contain highly useful information for the improvement of work organization, planning and overall efficiency. While models and techniques to extract these insights are well known and constantly developed in mathematical biology, complex systems physics and data science, the adoption of these techniques among Clinicians and Hospital Administrators remains limited, mainly due to language and methodological divides. Objective: In this article we implement an AI Multi Agent System (MAS) that allows Users to use predictive models and make statistical analyses via natural language conversations. Methods: We developed EpidemIA, a training-free MAS endowed with a modular suite of epidemiological models written by human experts, executable on a single-GPU workstation. To avoid relaying sensitive data to third parties, the system uses locally an open-weights LLM as backbone for its two subagents. Res
agentai agentagent system - arxiv:2610.39676 · cs.LGEvent-SAMA: Language-Grounded Video Object Segmentation on Neuromorphic Event StreamsSergei Fedorchenko, Max Zehnder, Nikola Zubić, Rong Zou +1
Event cameras offer microsecond temporal resolution, high dynamic range and no motion blur, but referring video object segmentation (RVOS) lacks event-stream datasets and models. We introduce Event-SAMA-239K, a synthetic-event conversion of the SAMA-239K grounded video-chat corpus and all additional SAMA training data: 30,650 video clips and 392,162 images. Optical-flow, FILM or saccade upsampling followed by a linear event generator preserves all mask, referring-expression and dialogue annotations. We also introduce Event-SAMA, adapting SAMA-4B with a two-channel polarity patch embedding initialised from pretrained RGB weights and a zero-initialised event adapter. End-to-end training yields 24.37 $\mathcal{J}\&\mathcal{F}$ on MeViS valid_u and 22.33 $\mathcal{J}\&\mathcal{F}$ on the converted ReVOS valid subset, establishing the first event-native RVOS baseline (RGB SAMA-4B: 55.4 on MeViS). Frozen RGB SAMA-4B with pseudo-RGB renderings or E2VID reconstructions of the same events score
event camera - arxiv:2610.39671 · cs.CLIncentivizing Self-Learning in Large Language Models via Reinforcement LearningLingkai Kong, Yuzhe Gu, Haiteng Zhao, Zijian Wu +5
Large language models (LLMs) draw on parametric knowledge and retrieval-augmented generation (RAG) to solve knowledge-intensive tasks. Yet problem solving in unfamiliar domains calls for tightly coupled knowledge acquisition and reasoning, beyond factual retrieval and aggregation or learning solely from a preassembled context. We study self-learning -- autonomously retrieving, studying, and applying knowledge from long-form reference materials on the fly. To investigate this capability, we formulate a textbook-based mathematical problem-solving task. LLMs must navigate textbook exposition, learn and apply relevant principles, and construct proofs for exercises, with reference solutions withheld. We introduce a BookNavigator interface to support this human-inspired study process and incentivize self-learning through reinforcement learning (RL) with fine-grained rewards. Experiments show that our method improves performance across in-domain and out-of-domain textbooks and subjects, raisi
long-contextretrieval-augmentedbenchmark - arxiv:2610.39665 · cs.LGA New AFIB Detection Method with Deep Neural NetworksSara Gjorgjieva, Ana Angjelevska, Dimitar Trajanov, Marjan Gusev
Early detection of atrial fibrillation (AFIB), characterized by irregular heart rhythms, can potentially save thousands of lives annually [20]. Numerous studies have attempted to address this challenge using various algorithms and methodologies. This paper introduces a novel and robust method that enhances the precision of AFIB detection. Neural Networks (NN) are widely recognized for their ability to identify complex electrocardiogram (ECG) patterns, facilitating the accurate detection of irregular heart rhythms. This study explores multiple conventional machine learning algorithms alongside neural network approaches to develop a robust AFIB detection system. To evaluate the effective- ness and robustness of the proposed method, the MIT BIH arrhythmia database (MITDB) and the Long-term Atrial Fibrillation Database (LTAFDB), both available on the Physionet site. Our approach presents a novel feature engineering technique, employing specific features such as a time series of the differe
benchmark - arxiv:2610.39664 · cs.CVEditAgent: Rethinking Agentic Image EditingHaiyang Xu, Tianyu Wang, Jialu Li, Jiteng Mu +4
Agentic image editors decompose complex requests into steps, but when a step fails, they mostly retry with the same narrow set of tools and a judge that scores each output by eye directly. We argue that such agents are bounded by two spaces: the action space of what the agent can do, and the observation space of what its judge can see. On our expert-annotated benchmark, a general-purpose editor yields no usable result in over 40% of the cases, and no judge is reliable across edit types. We present EditAgent, which expands both spaces. It routes each step to general editors and specialized methods, and attaches to each candidate a verification record. With these records, off-the-shelf VLM judges outperform finetuned reward models. Candidates are generated in parallel and judged in one listwise call: experts prefer this to sequential refinement in 78.5% of decided comparisons, at no more than half the judge calls. EditAgent improves two open-weight editors on three benchmarks and serves
agentagenticbenchmark - arxiv:2610.39660 · cs.MASticky Collusion: Persistence of LLM Agent Collusion after Incentive Removal and Partner ReplacementJoshua Adrian Cahyono
We ask whether collusion between AI agents survives once formed, when its incentive is removed or a colluder is replaced. Q-learning pricing agents stop colluding with a new rival; we test whether LLM agents do. We use the joint-task environment of Shi et al.\ (2026), where two agents verify each other's work under a rule they cannot satisfy and a reward that favours acceptance. We save each pair's history once its collusion has formed and resume it under interventions, with three models (GPT-6-Luna, DeepSeek V4.1 Flash and Gemini-3.1-Flash-Lite), \TotalHistories{} saved histories and \TotalRuns{} runs. Collusion is \emph{sticky}. First, it outlives its incentive: told that the reward has stopped, pairs still accept each other's unverified work in \GptToldRate, \DsToldRate{} and \GemToldRate{} of episodes, including pairs whose collusion began with an explicit agreement, and a payoff penalty has no significant effect. Second, replacing a colluder ends the pair's collusion but not the c
memoryagentai agentllm agent - arxiv:2610.39649 · cs.CVLook at Where You Think: Localization-Head Alignment for Fine-Grained Perception in Thinking VLMsYuan Wang, Wei Wu, Hongning Wang
Thinking vision-language models reason for thousands of tokens, yet still miss small details in high-resolution images. Zooming methods pay for extra inference passes, and outcome rewards never tell the model where to look at. In Qwen3.5-4B, we find that the bottleneck is selection, not seeing: a few attention heads, which we call localization heads, pick the evidence out of thousands of visual tokens; ablating four of them costs 13.2 points more than ablating four random heads, and their focus fades as the reasoning grows. We therefore propose Localization-Head Alignment (LHA), which makes attention a training signal: whenever the reasoning states a visual fact, these heads learn to look at the focal region rather than the background, as an auxiliary loss in both SFT and RL training. On benchmarks of V*, HR-Bench 4K and TreeBench, LHA lifts Qwen3.5-4B from 74.1% to 76.6% accuracy, the best single-pass 4B result we evaluate, ahead of multi-pass zooming and released 7B models. Alignment
benchmark - arxiv:2610.39646 · cs.CVFlexTry-On: A Time-Efficient Framework for Multi-Scenario Virtual Try-onPhuong Dam, Quyen Nguyen Duc, Son Hoang Minh, Daeyoung Kim
Virtual try-on has achieved remarkable progress in recent years, yet most existing methods are designed for a specific scenario, such as cloth-to-person or person-to-person try-on, and often require separate conditioning strategies and processing pipelines for different settings. We present FlexTry-On, a unified and time-efficient diffusion-based framework for diverse virtual try-on scenarios, including cloth-to-person, person-to-person, cross-domain, and in-the-wild settings. FlexTry-On adopts a common conditioning formulation, where a text-aware condition extraction strategy converts heterogeneous garment sources into a unified representation, and a dual-conditioning diffusion transformer jointly models garment appearance and person identity. Extensive experiments demonstrate competitive performance on both cloth-to-person and person-to-person benchmarks while maintaining fast inference. Moreover, the same framework can be directly applied to unseen cross-domain scenarios without req
benchmark - arxiv:2610.39641 · cs.AIRebuilding a Field's Open-Problem Catalog from the Literature: A Two-Layer Knowledge Graph, and Why Status Must Stay ProposedMadhulatha Mandarapu, Sandeep Kunkunuru
A research field records its open problems in surveys, keynotes and community reports. Those are prose, written every few years, and nobody can ask them who works on a problem now or whether it was settled last year. We argue the problems should be data, and we build the system that makes them so: a hand-curated problem layer carried in unchanged, under an evidence layer rebuilt from a bibliographic index, joined by typed, provenance-carrying links in one graph database and queryable from one language. A language model does four jobs -- per-paper extraction, catalog grounding, per-paper status verdict, and idea generation -- each constrained to a candidate set so out-of-set answers are refused rather than stored. We report a study on two curated catalogs, database research (1,053 problems) and AI research (1,000 problems), each with a prior catalog-only graph. The rebuilt graph answers every question the catalog-only graph answered -- 47 of 47 on databases, 34 of 35 on AI -- plus quest
knowledge graph - arxiv:2610.39639 · cs.CVWhere Do Vision Language Models Fail at 3D Reasoning?Khai Le-Duc, Nguyen Kim Hai Bui, An Ngo, Nam D. Hoang +10
Vision-language models can get individual judgments right while struggling to recover and consistently use the exact properties of a 3D object. To characterize this gap, we introduce BedBench, a controlled benchmark of thirteen procedurally generated tasks spanning individual judgments and composite answers. Across twelve open-weight and four frontier models, per-color counting exposes a pronounced weakness: humans achieve 93.3% accuracy, compared with 7.2% for the best model. Targeted fine-tuning raises one model's accuracy from 3.0% to 68.6%, but reveals a further difficulty: its answers about the most common colors remain inconsistent with its own per-color counts. Supplying those counts as text improves its judgments, showing that recovering a property does not mean using it consistently. Reasoning, scaling, and added visual cues each improve individual judgments, but none makes exact answers reliable. To locate the difficulty, we probe Qwen3.5-27B: exact per-color counts remain re
benchmark - arxiv:2610.39621 · cs.AIThe Real Cost of Context: Predicting Compute Time and Energy Consumption in LLM Prefill and Decode PhasesLars Krupp, Daniel Geißler, Francisco M. Calatrava-Nicolas, Stefan Gerd Fritsch +2
The integration of Large Language Models (LLMs) into agentic architectures has substantially increased token consumption, raising concerns about their energy footprint. Yet, estimating the energy consumption of proprietary API-based models remains challenging, and existing approaches focus primarily on generated output tokens while overlooking the impact of the large input contexts characteristic of agentic workflows. To address this gap, we systematically profile compute time and energy consumption across a wide range of prefill (input) and decode (output) token lengths. Using generalized linear models, we show that decode compute time and energy scale linearly with output length, whereas prefill compute time and energy exhibit both linear and quadratic scaling with input length. These findings demonstrate that input context becomes a relevant factor for large-context workloads. Furthermore, we show that compute time serves as a reliable proxy for energy consumption when direct energy
agentic - arxiv:2610.39620 · cs.AILetting AI Formulate the Research Question: Autonomous Research AgentsKo Fujimura
Systems such as The AI Scientist have shown that AI agents can carry out and write up research, but in those demonstrations humans chose the research area and selected the outputs to show. We study letting the AI formulate the research question itself, and we report every run. Our system gives an off-the-shelf coding agent an instruction document and two skills: one for formulating the question, in which the agent diverges from recent technologies into problems of other fields, screens its candidates for known failure modes and obtains a critique from a different model; and one for peer review, in which a reviewer from another model family reruns the study from its records before the agent revises the paper. Over 13 development runs, changes to the one or two sentences that set the research area coincided with large shifts in what the candidate questions and the chosen papers were about, which led us to the area of a new technology applied to a societal problem. We froze the system aft
agentai agent - arxiv:2610.39619 · cs.CVFutureWorld: A Plug-and-Play Causal Next-State Predictor for Interactive Video World ExplorationGuangyuan Li, Kelu Yao, Shuolin Xu, Xinlei Yu +5
Interactive video world models aim to follow user actions while maintaining scene consistency over extended interactions. History-based methods reuse generated latents to preserve scene content, but three limitations of this reuse contribute to error recirculation. 1) The latent generated at each interaction is retained as history, potentially containing blurred details, geometric distortions, or semantic drift. 2) Although retrieved historical content matches the current scene, it can still contain errors that are carried into the next chunk. 3) New outputs then become history for later interactions, potentially reinforcing the same errors through repeated reuse. To limit this cycle, we aim to predict the future scene state from the current generated latent and the next action before generating the next interaction's video chunk. This future-state prior provides guidance for the next chunk, with the aim of reducing the errors carried forward from history. To this end, we propose a plu
world modelaction-conditionedbenchmark - arxiv:2610.39617 · cs.CVUniTele: Learning Device-Conditioned Degradations for Unified Smartphone Telephoto Super-ResolutionYuchen Zhang, Aojie Li, Wei Cheng, Xianfang Zeng +4
Super-telephoto imaging on mobile phones suffers from complex, device-dependent degradations that are poorly captured by conventional handcrafted pipelines. Collecting large-scale aligned training pairs across different phones and diverse scenes, however, is both costly and difficult to scale. We therefore study whether device-specific degradations can be learned from only a small number of physical captures and transferred for training a unified super-resolution (SR) model. Our key idea is to factor degradation learning into two learnable components: one captures the device-specific imaging characteristics under well-resolved conditions, while the other captures the additional information loss and artifacts that arise in super-telephoto imaging. We build a paired benchmark across three phone brands (Huawei, Apple, and vivo) by photographing the same content at two display sizes and zoom settings. The larger displayed image is captured at the phone's maximum optical zoom setting, while
benchmark - arxiv:2610.39609 · cs.CVDiFuse: Fusing Native Diffusion Representations for Spatial Reasoning in Unified Multimodal ModelsRuixin Fan, Weining Wang, Jiahui Sun, Tailai Chen +1
Unified multimodal models (UMMs) jointly support visual understanding and image generation, yet information flow between these two capabilities remains largely asymmetric: understanding representations routinely condition generation, whereas generative representations are rarely reused for multimodal reasoning. In this work, we investigate whether intermediate representations from a UMM's native diffusion branch can enhance spatial reasoning without compromising general multimodal understanding. We propose DiFuse, which selectively incorporates intermediate diffusion representations into the understanding pathway of UMMs. Through systematic layer-wise analysis, we identify which diffusion representations are most informative and where they can be integrated with minimal interference to semantic processing. Guided by these findings, DiFuse spatially aligns diffusion features with visual tokens and injects them into selected multimodal decoder layers via residual fusion. Extensive experi
benchmark - arxiv:2610.39608 · cs.AIFrom Intent to Improvement: Agentic Dataset Search for Better Model TrainingQiaosheng Chen, Kaijia Huang, Fanqi Zhou, Gong Cheng
Finding training data in million-scale machine learning repositories is a Web search problem where text relevance is not enough. A query couples a semantic training goal with constraints such as task, license, and training stage, and the metadata that would verify them are often missing. Text retrieval ranks dataset cards by similarity but cannot enforce hard constraints or follow the relations that reach a sparsely described dataset. Free-form tool-calling agents are more flexible, but the same model that picks a dataset also declares it a match, and nothing checks that the query's constraints are actually met. Neither approach establishes that the data it returns will improve model training. To address it, we propose AgenticDS, a contract-governed agentic framework in which every language model decision instantiates a typed object before deterministic code consumes it. The framework compiles a query into a typed search plan, executes it through admissible deterministic operators over
llm agentagenticbenchmark - arxiv:2610.39602 · cs.CVEvaluating Cultural Knowledge Editing in Vision-Language ModelsMalvina Nikandrou, Georgios Pantazopoulos, Ioannis Konstas, Alessandro Suglia
Current Vision-Language Models (VLMs) frequently fail to recognize culture-specific concepts, especially from underrepresented regions. Effectively injecting this long-tail knowledge into existing models remains a challenge. In this work, we investigate knowledge editing as a targeted alternative by introducing CulturalBind, a benchmark comprising 3.2K images of 1K concepts from 6 domains, designed to evaluate culture-aware cross-modal knowledge as well as concept-level knowledge editing in VLMs. Pre-edit evaluation reveals that textual and visual knowledge is decoupled: models often demonstrate text-based knowledge of cultural concepts but fail to identify them from images. We then evaluate editing methods in single and sequential scenarios. Although current methods successfully correct individual visual recognition errors, they exhibit low cross-modal portability, failing to integrate edits into the model's broader parametric knowledge. Furthermore, performance degrades rapidly under
benchmark - arxiv:2610.39585 · cs.CVTactoHand: Learning Geometry-Based Dense Hand-Surface Interaction Fields from VideoZiheng Xi, Weibo Li, Jiayi Yu, Sipeng Zheng +5
Understanding how humans interact with the world requires a representation of physical interactions at the hand surface. Tactile sensing captures these interactions, but collecting tactile data at scale requires specialized hardware. We introduce TactoHand, a model that predicts three dense hand-surface interaction fields---contact, proximity, and 3D force---from monocular video. We derive geometry-based contact and proximity labels and motion-derived force labels from eligible training data within a curated collection spanning approximately 50 million source frames. TactoHand uses a temporal transformer to aggregate visual evidence across frames, while three learned queries decode the corresponding hand-surface fields. A differentiable physics layer further constrains force predictions using observed object motion. Across in-domain and out-of-domain benchmarks, TactoHand improves contact and proximity estimation over prior methods. In addition, physics supervision reduces position and
tactilebenchmark - arxiv:2610.39582 · cs.ROKinematic-Tactile Aligned Dexterous Retargeting with Contact Wrench Space AlignmentXinyuan He, Ziheng Xu, Yueyuan Chen, Yufan Ye +5
Human demonstrations provide a natural and scalable source of data for learning dexterous manipulation policies. However, existing retargeting methods primarily transfer human kinematics, while the tactile information underlying physical interaction is rarely preserved across embodiments. To address this limitation, we present a hierarchical three-stage framework for kinematic-tactile aligned dexterous retargeting. We use object-centric contact wrench space as a physical representation for tactile alignment between human and robot hands. The framework first aligns hand kinematics and contact geometry, and then constructs a human wrench support profile from tactile forces and contact geometry. Robot fingertip force targets are subsequently optimized to cover the human wrench support and converted into joint position corrections, yielding directly replayable robot hand trajectories. We evaluate the framework in simulation and real robot experiments across three pouring conditions. In the
manipulationdexteroustactile - arxiv:2610.39576 · cs.MABayesian Offline Multi-Agent Reinforcement Learning with Latent World ModelsXi Xiong, Lu Liu, Yiming Li, Jakob Foerster
Model-based offline multi-agent reinforcement learning (MARL) provides a promising approach to multi-agent decision-making under partial observability, where the global state is unavailable from offline data. However, limited dataset coverage can lead to substantial out-of-distribution (OOD) model errors, which may mislead policy optimization and degrade downstream performance. Existing model-based offline MARL methods typically rely on access to global states or mitigate model errors through pessimistic or conservative policy optimization, leaving the reliability and refinement of learned world models less explored. We propose a Bayesian offline MARL framework that infers latent global states from decentralized observation-action histories and iteratively refines a posterior world model using only offline data. Building on posterior information loss, we extend this criterion to the multi-agent setting as multi-agent posterior information loss (MPIL), which provides a reliability measu
world modelmulti-agent - arxiv:2610.39570 · cs.AILLM Agent-Based Design of RISC-V CPUs with FPGA-in-the-Loop EvaluationLeo Itoh, Junichiro Kadomoto
Long-run RTL simulation of CPU candidates can take hours, while short diagnostic runs may miss performance degradation that emerges later in execution. This evaluation cost can delay design decisions in LLM-agent-based CPU development. We present a development environment in which LLM agents implement CPU candidates on an FPGA and use on-board performance measurements to guide design decisions. We applied it to the human-supervised development of a single-cycle RV32I core into a three-wide out-of-order RV32IM core. The final design reaches 5.61 CoreMark/MHz at 130 MHz on a Artix-7 FPGA. For the same CPU candidate running 8,000 CoreMark iterations, FPGA evaluation, including implementation, was 9.6x faster than single-threaded RTL simulation. In one case, board measurements revealed long-run performance degradation in a candidate that improved in a short diagnostic run. The result guided candidate selection and led to RTL-based root-cause analysis and a fix.
llm agent - arxiv:2610.39563 · cs.LGTrajeform: Embedding Crowd DataMarilena Lemonari, Panayiotis Charalambous, Yiorgos Chrysanthou
Crowd simulation aims to generate agent behaviours that are rich, natural, and appropriate to their surrounding environment. Recent advances in Machine Learning have enabled such behavioural patterns to be learned directly from trajectory data, allowing simulation models to capture increasingly complex crowd dynamics. However, the capabilities of these systems remain strongly dependent on the type and diversity of the data used during training. When evaluating how well crowd simulators generalise to real-world datasets, most existing metrics focus primarily on trajectory-level differences rather than behaviour-level similarity. In this paper, we present Trajeform, a framework that trains a reconstruction VAE (Variational Autoencoder) on real trajectory data to learn a meaningful continuous embedding that preserves behavioural characteristics of real crowds. The trained encoder can then be applied to synthetic trajectories generated by simulators and used to compare the resulting embedd
agent - arxiv:2610.39560 · cs.ROSimTO-Robust: Automated Design of Soft Grippers Robust to Grasp Pose and Object GeometryKurt Enkera, Joshua Pinskier, Marcus Gallagher, David Howard
Robotic grippers operating in unstructured environments must robustly grasp objects of different shapes and sizes despite perception and positioning errors. Soft grippers that can achieve this robustness through intelligent mechanical design rather than sophisticated control software offer a promising solution. However, no existing computational design framework explicitly optimises soft gripper morphology for robustness across object geometries and grasp poses. We introduce SimTO-Robust, a generalisation of simulation-driven topology optimisation (SimTO) that simulates grasps across multiple objects and poses and aggregates their contact forces to guide gripper morphology optimisation. We generate 6,646 designs and physically evaluate 24 of them alongside 10 specialised SimTO designs, two rigid baseline grippers, and a baseline generalist soft gripper produced using conventional topology optimisation. All 37 grippers are tested in a pick-and-place task involving five feature-rich obje
grippergrasp - arxiv:2610.39558 · cs.LGTabCUQ: Context-Uncertainty Quantification for Tabular Foundation ModelsYuebei Shi, Molei Liu, Zijian Guo
Tabular foundation models use in-context learning to estimate conditional expected outcomes and average treatment effects for specific groups of individuals. Using these estimates to guide scientific decisions requires reliable uncertainty quantification. The challenge is to account simultaneously for two sources of uncertainty: limited-context bias and context variation arising from the small labeled sample supplied at inference. Predictive spread does not directly quantify both sources of uncertainty, and we observe that the ordinary bootstrap may substantially overestimate context variation, whereas subsampling may underestimate it. We introduce TabCUQ, a perturbation-based framework that injects noise into observed outcomes and applies the foundation model to the perturbed contexts. TabCUQ estimates context variation using wild-bootstrap dispersion and accounts for limited-context bias using the maximum deviation between perturbed and original estimates. For a kernel-regression ana
benchmark - arxiv:2610.39557 · cs.LGLatentWeave: Test-Time Latent Optimization for Multimodal ReasoningFanci Meng, Kun Wang, Guibin Zhang, Xinlei Yu +3
Recent advances in test-time optimization have shown that the multimodal reasoning performance of large multimodal language models (MLLMs) can be improved during inference without modifying backbone model parameters. However, existing approaches primarily operate at the prompt or output level, leaving the latent reasoning process largely uncontrolled and insufficient to prevent error accumulation during autoregressive generation. To address this limitation, we propose LatentWeave, a test-time framework that treats multimodal reasoning as a controllable latent trajectory. LatentWeave performs structured Latent Weaving, progressively refining intermediate hidden states segment-wise to correct emerging deviations, and further introduces Reasoning Anchors via lightweight WEAVE tokens to stabilize reasoning structure and guide answer-oriented generation. These mechanisms are designed to enable fine-grained, step-aware intervention in the latent reasoning process while keeping the main MLLM
benchmark - arxiv:2610.39556 · cs.LGExplainable and Adaptive Machine Learning for Fraud Risk Detection in Bangladesh's Digital Banking EcosystemMd Najmus Swaqeeb
The rapid expansion of digital banking and mobile financial services creates growing challenges for fraud detection, particularly under severe class imbalance and operational constraints. This study presents an explainable and adaptive machine-learning framework integrating chronological evaluation, leakage-conscious feature engineering, probability calibration, SHAP-based explainability, and adaptive risk prioritization. Experiments use the synthetic PaySim mobile-money benchmark and compare Logistic Regression, Random Forest, XGBoost, and LightGBM. XGBoost achieved the strongest fixed-threshold performance, with PR-AUC of 0.5658, ROC-AUC of 0.9068, and F1-score of 0.5369. Platt calibration reduced the test Brier score from 0.1141 to 0.0245, improving probability reliability. SHAP analysis identified temporal, destination-account, transaction-type, and amount features as important model drivers. An adaptive top-1% policy reduced false positives from 192 to 117, but recall declined fro
benchmark - arxiv:2610.39547 · cs.ROAssuring Learning-Enabled Cyber-Physical Systems with Foundation Models: A Neuro-Symbolic Agenda for Testing, Verification, and Runtime AssuranceXi Zheng, Ivan Ruchkin, Hong Jin Kang, Tevfik Bultan +10
Learning-enabled cyber-physical systems (CPS), including autonomous vehicles and UAVs, increasingly incorporate foundation models, yet fragmented assurance practices provide insufficient support for system-level guarantees. We argue that restoring engineering control across stochastic neural, symbolic, and physical boundaries requires a shift from model-centric evaluation to system-level assurance. We introduce the Foundation Model Assurance Stack, connecting three directions: (i) neuro-symbolic architectures that constrain foundation models through explicit symbolic world models, reasoning layers, and interfaces; (ii) semantically grounded, conformance-based testing that combines formal specifications and symbolic world models with foundation-model-assisted specification mining, scenario generation, and test synthesis; and (iii) compositional verification and runtime assurance that target formally tractable components and interfaces through modular contracts, equivalence checking, ver
world model - arxiv:2610.39543 · cs.ROFrom Reasoning to Action: Consequence-Grounded Alignment for Vision-Language-Action ModelsYixue Zhang, Wenke Xia, Kun Wu, Wenbo Yu +5
Reasoning informs decision-making by connecting intended actions with their expected consequences. Recent vision-language-action (VLA) models incorporate embodied chains of thought (ECoTs) to guide action generation. However, such guidance remains implicit, as actions are conditioned on ECoTs rather than explicitly optimized to align with them. We introduce CoRAct, a consequence-grounded reasoning-to-action alignment framework that explicitly aligns reasoning and action through reward-guided optimization in action noise space. CoRAct formulates two physically grounded ECoT components as verifiable signals: movement specifies intended motion, while waypoints encode expected spatial consequences. A unified Mixture-of-Transformers jointly models ECoT reasoning, action generation, and action-conditioned future prediction. For candidate actions generated from sampled noise, CoRAct computes alignment rewards based on movement consistency and agreement between predicted outcomes and specified
vision-language-actionvlaembodiedmanipulationaction-conditioned - arxiv:2610.39540 · cs.LGThe Cost of Working Memory in Streaming Mean EstimationZeyu Wu, Zihao Wang, Minghua Deng
A streaming estimator must compute from the evidence that fits in memory. We determine when storing an accurate mean is easier than computing it. For Gaussian mean estimation over the Euclidean unit ball in $\mathbb{R}^d$, we establish the minimax rate under squared error with $K$ persistent states, a known horizon, and a public time index: $\min\{1, K^{-2/d}+\frac{d}{n} (1+\frac{\log(ed)}{1+(\log K)/d})\}$. Unrestricted computation followed by $K$-level encoding has rate $\min\{1,d/n+K^{-2/d}\}$. At a fixed sufficiently small error, statistically optimal sample size therefore requires order $d\log d$ working bits, while order $d$ bits suffice to store the answer. The lower bound tracks repeated state compression through an entropy potential obtained by Gaussian smoothing. A matching construction combines a multiscale modular counter with workspace reused across coordinate groups. In one dimension, it attains the terminal rate $n^{-1}+K^{-2}$. Any fixed confidence requirement that the
memorypersistent state - arxiv:2610.39532 · cs.AIJudging Before Propagation: Task-Graph Contextual Integrity for Multi-Agent PrivacyHengyu Wu, Blaise Delattre, Yang Cao
Multi-agent LLM systems increasingly rely on persistent memory and inter-agent communication. These capabilities create privacy risks that extend beyond individual interactions: a release appropriate in isolation may become inappropriate when combined with information previously disclosed to a recipient or propagated to downstream agents along the planned workflow. Existing Contextual Integrity (CI) mechanisms for agentic systems typically rely on local judgments that assess each disclosure in isolation, with or without regard to the recipient's history, but do not account for how information propagates downstream. However, compliance at the current boundary does not guarantee appropriateness across the workflow. We introduce Task-Graph Contextual Integrity (TG-CI), to our knowledge the first task-graph-native formalization of Contextual Integrity for multi-agent systems. TG-CI represents a workflow as a Multi-Agent Task Graph (MATG) and extends CI from atomic transfers to stateful, gr
memorypersistent memorymulti-agentagenticagent systembenchmark - arxiv:2610.39531 · cs.LGConic-Lifting for Learning-rate-free Convex OptimizationJulien Grand-Clément, Joon Kwon, Francesco Orabona
A central topic in online and stochastic optimization is to design algorithms that do not require tuning of hyperparameters. Successful strategies in online learning include AdaGrad-style learning rates, Conic Blackwell algorithms, and parameter-free algorithms. However, these classes require different, often difficult, analyses. We develop a unifying family of Online Convex Optimization (OCO) algorithms that require no user-tuned learning rate on convex domains that may be unbounded. The construction lifts the domain to a cone by adding one scalar coordinate, an idea going back to Blackwell (1954). The lifted subgradients are orthogonal to the corresponding lifted decisions, leaving the comparator scale free to optimize in the regret analysis. Applying Follow-the-Regularized-Leader (FTRL) and Online Mirror Descent (OMD) on the cone recovers Conic Blackwell Algorithm (CBA) and Conic Blackwell Algorithm Plus (CBA$^+$) on bounded domains in Hilbert spaces and extends them to finite-dimen
online learning - arxiv:2610.39530 · cs.CVMorphSIG: Subject-Driven Image Generation via Decoupled Anchoring and Feature TransportHailong Yan, Yongrui Zhang, Xiangtao Zhang, Le Zhang
We propose MorphSIG, a training-free framework that reformulates subject-driven image generation (SIG) as pseudo-video feature transport. MorphSIG contains two core modules: Progressive Decoupled Anchoring, which uses frequency-domain decoupling and structural scrambling during early denoising to break structural locking and enable a gradual transition from free composition to identity alignment; and Pseudo-flow Guided Feature Transport, which constructs a latent-space pseudo-flow field to mimic Image-to-Video temporal coherence and transport subject semantics precisely. To address stylistic homogeneity and the lack of ground truth in existing benchmarks, we further introduce a benchmark with diverse stylized subjects. Extensive experiments show that MorphSIG significantly improves pose diversity and subject fidelity over baselines without additional training. Overall, MorphSIG bridges static generation and dynamic propagation, providing a cost-effective paradigm for high-fidelity subj
benchmark - arxiv:2610.39528 · cs.ROPRIME W0: Physical Priors from World Models for Embodied ManipulationHuabin Peng, Fagang Jin, Yancong Wei, Tengfei Ma +3
Vision-Language-Action (VLA) models excel at instruction following, yet most remain feed-forward: they condition actions only on the current observation and discard both long-horizon temporal context and the hierarchical structure of VLM representations. We present PRIME W0, a hierarchical conditional flow-matching framework that closes this gap through three synergistic designs: (1)per-layer KV injection from a pretrained VLM into an Adaptive MoE action expert; (2)dynamics-aware conditioning via asynchronous world-model subgoal-image prediction and a DINOv2 historical-frame semantic bridge; and (3)Action Dynamic Manifold (ADM) flow matching with covariance-structured noise that respects robot kinematic coupling. PRIME W0 is trained in four stages---VLM pre-training, DAT/NTP visuomotor pre-training, DiT flow-matching decoder assembly with WM injection, and end-to-end SFT---and achieves state-of-the-art results on LIBERO and real-robot benchmarks, with systematic ablations validating ea
vision-language-actionembodiedmanipulationliberoworld modelbenchmark - arxiv:2610.39517 · cs.CVFFAI: Flow Matching with Feedback Training and Adaptive Initialization for CT ReconstructionJigang Duan, Heran Wang, Genwei Ma, Xing Zhao
Although low-dose, sparse-view, and limited-angle CT involve distinct degradation mechanisms, all three pose reconstruction challenges due to limited measurement information and can therefore be addressed within a unified framework. In iterative image restoration with flow matching, prediction errors may cause inference states to deviate from the prescribed training path, while a fixed initialization may be unsuitable for inputs with varying degradation severity. We propose a unified conditional flow matching method that conditions on the initial reconstruction and combines feedback-aware training with degradation-adaptive initialization. During training, a teacher network simulates inference updates to generate shifted states, from which a student network learns to recover high-quality reference images, improving its ability to handle prediction errors. During inference, the starting time is selected based on the noise level or extent of missing projection data, and the noise added to
benchmark - arxiv:2610.39516 · cs.CVRobust LIDAR Sensor-to-Sensor Domain Generalization via Accumulated-Point Scene CompletionChristoph B Rist, Larissa T Triess, Markus Enzweiler, Dariu M Gavrila
We study domain generalization for LIDAR semantic segmentation. Deep neural networks achieve strong performance on 3D point clouds, yet real-world LIDAR exhibits sensor-specific sampling patterns that hinder transfer across datasets. In many applications, performance on completely unseen sensors is crucial. A model should preferably work on a new setup without prior access to its data or labels. To this end, we introduce a source-only training scheme that adds an occupancy prediction auxiliary objective. We aggregate source-domain scans into a simple world model and use it as the supervision signal, replacing single-scan standard segmentation training. A key property of our approach is that it does not alter the input scan. During training and testing the network consumes the same single-frame LIDAR input, only the supervision signal is enriched via the occupancy objective. At deployment, there is no runtime overhead as the auxiliary head can be dropped for pure segmentation, or kept t
world model - arxiv:2610.39508 · cs.AICompile Away Uncertainty: Self-Evolving Agent Skills via Programmatic CompilationChengyu Yang, Yunxiang Zheng, Zongzheng Li, Jiamei Li +5
Self-evolving agent skills enable AI agents to accumulate and refine reusable procedural knowledge from interaction experience without updating model parameters. However, since such skills are typically represented as textual documents and executed through LLM reasoning, even previously learned procedures must be repeatedly interpreted and regenerated, introducing avoidable uncertainty into every execution. In contrast, program execution is deterministic and independently verifiable. Intuitively, transforming well-specified operations into executable programs can reduce reliance on LLM reasoning and, consequently, reduce execution uncertainty. Motivated by this intuition, we introduce SKillCompiler, a self-evolving skill framework that integrates programmatic compilation into the skill evolution process. Specifically, SkillCompiler adopts an executability-aware skill compilation strategy that decomposes skills into fine-grained operations, analyzes their executability, and constructs h
agentai agentself-evolvingbenchmark - arxiv:2610.39506 · cs.LGInterpretable by Design Tabular Prediction via In-Context Local Linear LearningRan Eisenberg, Steven Berger, Ofir Lindenbaum
In-context learning (ICL) lets tabular foundation models make predictions without dataset-specific training or fine-tuning. These models achieve strong predictive accuracy, but interpreting how individual features contribute to the prediction remains difficult. In biology, for example, feature contributions can help identify biomarkers or risk genes. Post-hoc interpretability methods assign feature importance scores, but these scores do not indicate what the model actually uses and can lead to explanation errors. We introduce TICL$^3$ (Tabular In-Context Local Linear Learning), an interpretable-by-design predictor that uses a tabular foundation model to guide a simple linear prediction rule. TICL$^3$ first learns a shared linear model to capture dataset-level feature effects. Then, for each query, it retrieves similar training samples using representations from the foundation model and fits a second linear model to the residuals in that neighborhood. The shared model captures broad eff
benchmark - arxiv:2610.39505 · cs.CVAgent-Wise Test-Time Normalization for Camera-Based Cooperative Perception with Corrupted CollaboratorsYu Wang, Shaohua Wang, Yicheng Li, Kecheng Sun
Intermediate-fusion cooperative perception lets a vehicle detect objects beyond its own field of view by fusing features shared by nearby agents. When only a collaborator's cameras are degraded by fog, motion blur or exposure errors, the features it shares drift away from the training distribution while the ego input stays clean, and cooperation can become worse than not cooperating at all. Test-time batch-normalization adaptation, the standard training-free remedy, estimates statistics over the pooled batch of all agents and therefore normalizes both the clean ego and the corrupted collaborator with a mixture of two distributions. We show that this mixing harms even the clean ego: when the ego shares BN statistics with a corrupted collaborator, its detection from its own features drops by 3.58 to 6.68 [email protected] points although its cameras are untouched. We propose AgentNorm, an agent-wise test-time normalization that i) estimates the normalization statistics of every collaborator from it
agent - arxiv:2610.39499 · cs.AIIntegrated Feature Selection for classification and Representation Learning with Group-Lasso Siamese NetworksNisargo Ranjan Pal, Pradeep Singh
The Least Absolute Shrinkage and Selection Operator (Lasso) has been widely employed for feature selection in linear models and has also been extended to multi-layer perceptron (MLP) networks for both feature and group feature selection. Group-Lasso has also been used in feature selection using feed-forward neural networks. To the best of our knowledge, this is the first study to incorporate the Group-Lasso principle into a Siamese network architecture for feature selection. Unlike existing approaches that primarily optimize classification accuracy, the proposed method aims to identify a compact subset of features that simultaneously preserves the intrinsic representation of the input data (the goal of Siamese network) and is expected to achieve high classification performance. Two variants of the proposed framework are developed: one based on the Euclidean distance and the other on a modified Jaccard similarity measure for real-valued vectors. The latter is specifically designed to im
benchmark - arxiv:2610.39497 · cs.AIDoes Scaling Toward AGI Depend on What System-1 Lets System-2 Execute?Qi Wang, Yun Qu, Lizhou Cai, Heming Zou +9
Progress toward artificial general intelligence (AGI) has largely relied on scaling a resource-intensive executive, System-2, through larger models, more data and compute, longer reasoning, and more agents. We ask whether the next stage may come instead from co-designing it with System-1, a lightweight, semantically aware predictor that evaluates candidate interactions before execution, so that System-2 acts only on those worth their cost. Our answer is a qualified yes. We study \textit{conditional interaction}, a predict--gate--execute--update loop: System-1 estimates each candidate's utility, System-2 executes those admitted by an adaptive threshold, and the outcomes update System-1. The loop spans training, test-time reasoning, adaptive decision-making, and agentic tool use, with utility and cost defined by the application. Under ideal screening and equal execution costs, execution compute falls by a factor approaching $1/ρ_u$, where $ρ_u$ is the density of utility-relevant interact
agentictool use - arxiv:2610.39477 · cs.LGTopology Survives Shuffling: Robust Long-Context Memory Access Prediction for Graph AnalyticsDongyan Sun, Neelesh Gupta, Rajgopal Kannan, Viktor K. Prasanna
Graph analytics generates irregular, topology-dependent memory-access streams via indirect traversals, limiting data prefetching. Accurate prediction enables timely prefetches that hide memory latency. However, ML predictors couple short histories with fixed horizons, fitting the training trace's node layout rather than learning transferable graph patterns. We propose LoCoMAP, a long-context representation-learning framework for memory-access prediction with independently controlled observation history and prediction horizon. We evaluate Connected Components (CC), PageRank (PR), and Single-Source Shortest Paths (SSSP) on six synthetic graph families, with Erdos-Renyi (ER) and Grid as random and deterministic controls. We test two interventions: same-family regeneration changes connectivity, while node shuffling disrupts layout but preserves connectivity. Across 18 graph-workload pairs, extending history from 8 to 128 accesses improves F1 by 7.8%, while extending prediction horizon from
memorylong-contextlong context - arxiv:2610.39468 · cs.AISoei: Structured Observability of Evidence and Intent for Multi-Agent AI Software EngineeringMohammed Nayeem Khan, K S Sarun Kumar
AI coding agents now do long, multi-step and multi-session software work with little supervision. When a developer comes back to that work, the useful question is not what the model said but what was actually done, by whom, and whether it is finished. We study this in three parts. First, we verify Agenttrail, an open-source local observability tool for coding agents, against its source at commit e4ba2da: nine documented constants and mechanisms hold at file-and-line level, two event pipelines are never unified, and its Map service has an unauthenticated local mutation surface. Second, we run a prototype of our system on 45 real coding-agent sessions (102,233 records, 20,577 tool executions, 2,336 human-to-agent episodes). With a blind-audited classifier (precision 83.3% and 96.7%) we estimate that 23.4% (95% CI 17.1 to 35.7) of code-editing episodes end with no independent check after the last edit, and that 57.8% of to-dos the agent marked complete after editing had no proof in betwee
agentmulti-agent - arxiv:2610.39467 · cs.CVCalibrating Attention-Based Token Pruning for Efficient Multi-Image Synthesis in UMMsZimeng Wu, Qijian Tian, Penghui Du, Changqian Yu +2
Unified Multimodal Models (UMMs) enable powerful multi-image synthesis but incur substantial computation from long visual contexts. While token pruning offers a direct route to accelerating inference, existing methods are primarily adapted from understanding tasks, suffering from insufficient reference image exploration, confused conditional guidance, and pooling-induced ambiguity that degrades generation quality. To address these issues, we propose CATP, the first training-free context-compression method tailored to UMM generation, which \textbf{C}alibrates the discrepancies of general \textbf{A}ttention-based \textbf{T}oken \textbf{P}runing for better quality-efficiency trade-off. Concretely, building on independent layer-wise pruning during denoising, we introduce Guidance-Consistent Pruning (GCP) to synchronize pruning decisions across conditioning branches, and a Taylor-Calibrated Attention Metric (TCAM) that derives a theoretically grounded second-order correction for pooling-ind
benchmark - arxiv:2610.39466 · cs.LGCBE: A Complexity-adaptive Boundary-aware Ensemble Method for Imbalanced DataXun Wei
In many real-world classification tasks, datasets exhibit class imbalance, which can degrade the generalization performance of conventional learning algorithms. Data complexity has recently been recognized as a more objective indicator of learning difficulty than the imbalance ratio alone, because class overlap and complex decision boundaries often impair classifiers more severely than imbalance itself. Samples near the decision boundary are inherently difficult to classify, which has motivated many researchers to focus on such instances to improve classifier performance. However, most under-sampling approaches fail to differentiate their sampling of boundary samples across datasets of varying complexity. To overcome this limitation, we propose a complexity-adaptive boundary-aware ensemble (CBE) framework that links the sampling weight distribution to data complexity and dynamically adjusts the boundary preference of under-sampling accordingly. We also introduce an overall complexity m
benchmark - arxiv:2610.39455 · cs.LGSparse Fine-Tuning for Parameter-Efficient Adversarial TrainingZhaoxin Wang, Zihang Ding, Handing Wang
Deep neural networks achieve strong performance across many tasks but remain vulnerable to adversarial perturbations. Adversarial training (AT) is one of the most effective defenses, yet it suffers from high memory cost on large models and the persistent clean-robust accuracy trade-off. Recent work introduces parameter-efficient fine-tuning, such as LoRA, into AT, but this constraint induces a notable robustness gap relative to full-parameter AT. In this work, we revisit parameter-efficient adversarial fine-tuning from a parameter space perspective. Through gradient analysis, we find that adversarial optimization is highly concentrated on a small subset of parameters, which we call Robustness-Critical Parameters (RCPs), suggesting that robustness is encoded in a sparse subspace. Building on this observation, we propose Sparse Adversarial Fine-Tuning (SAFT), which identifies RCPs via adversarial gradient saliency and performs AT by updating only these parameters while freezing the rest.
memory - arxiv:2610.39454 · cs.ROKairos: Semantics-Aware Orchestration of Planning Engines across Classical, Numeric, Temporal and Hybrid PDDLFrancisco Martín Rico, Esther Aguado González, Francisco Miguel Moreno Olivo, Rodrigo Pérez Rodríguez +2
Planning at run time, as in robots that re-plan while they act, needs a planner that reads the PDDL its users write (durative actions, numeric resources, time windows, continuous change), answers quickly, and delivers only validated plans. Specialised planners are strong on individual fragments of PDDL, but they differ in coverage, latency and in what their failures mean. We study whether selecting heterogeneous engines from the semantics of each task can give broad coverage of PDDL, from STRIPS to PDDL+, while remaining competitive with specialised planners and making explicit what each answer guarantees. Kairos profiles each task to decide which of five engines apply (ground heuristic search, compressed and snap-action temporal search, time steps for hybrid dynamics, lifted search) and schedules them under one budget of time and memory. Every candidate plan is checked by a validator independent of the planner's transformations, and "unsolvable" is reported only when a search whose ex
benchmark - arxiv:2610.39451 · cs.ROSmooth Action Chunking Via Movement-Primitive Representations for Imitation LearningShota Nomura, Takayuki Osa, Yasuhiro Kato, Kazuki Ota +2
Imitation learning has become a widely adopted approach for vision-based robot policies, owing to its conceptual simplicity, scalability, and ability to leverage large-scale datasets. As these policies have grown more capable, action chunking has become their standard interface. At each decision step the policy outputs several future actions at once. Backbones from diffusion policies to large vision-language-action models all share it. Whatever backbone produces the chunk, its temporal structure is learned only implicitly. The commanded trajectory is therefore jerky both within a chunk and between successive chunks. On real hardware this appears as visible vibration and as repeated saturation of the follower's velocity limit. We address this issue by incorporating a movement-primitive-based trajectory representation into action chunking. Specifically, a trajectory segment is represented as a linear combination of smooth basis functions, yielding smooth trajectories by construction whil
vision-language-actionaction chunking - arxiv:2610.39446 · cs.ROAsk to Evolve: A Unified Harness for Lifelong Embodied Navigation with Hybrid Action Space and Persistent MemoryKai Sheng, Liuyi Wang, Jinlong Li, Haojie Dai +2
Giving embodied agents the freedom to act autonomously, initiate interactions, and understand their surroundings through experience is a step toward bringing them to life in the physical world. Realizing this vision requires action, interaction, and memory to work together, so that each encounter can inform current decisions and contribute experience for future tasks. We present EvoLife, a unified navigation harness that enables a multimodal embodied agent to choose how to move, when to ask, and what experience to carry forward. Its hybrid action space supports waypoint navigation, local discrete adjustments, pixel-based navigation, and object-directed navigation. Active interaction allows the agent to seek clarification or assistance when it encounters ambiguity or difficulty, while persistent spatial memory binds observations and dialogue to places, retaining useful experience to guide subsequent visits. We evaluate eight benchmarks across four task families: vision-language navigati
embodiedmemorypersistent memoryagentembodied agentbenchmark - arxiv:2610.39441 · cs.CVPathCowork: Learning from real-world pathology practice to build AI copilotsSheng Wang, Jialu Yao, Maria V. Rossi, Shunsuke Koga +21
Medical AI learns from the answers doctors record, but rarely from how they interpret the evidence to reach them. Here we introduce PathCowork, a framework that turns pathologists' narrated examinations into supervision for visual grounding and procedures that guide diagnostic AI agents. Synchronized gaze, narration and navigation link visual findings to the process of interpreting them. Across 48 sessions from 16 pathologists, the platform captured 1,028 minutes of slide review. Using these data, we trained PathCowork-1T for server deployment and PathCowork-27B for local deployment, with improved visual grounding. Across five organ-specific evaluations, agents combining this grounding with recorded diagnostic procedures achieved primary task scores averaging 17.9 and 24.1 percentage points higher than Claude Code and Codex, respectively, with shorter processing times and lower inference costs. Compared to the state-of-the-art multiple-instance learning AIs, the PathCowork agent-propos
ai agentbenchmark - arxiv:2610.39439 · cs.AITimbre: Reliability-aware harmonic consensus for label-free multimodal test-time rexinforcement learningHaotian Lyu, Haoyu Wang, Ru Wang, Feng Ji +3
Label-free test-time reinforcement learning adapts vision--language models using rewards derived from their own predictions. However, misinterpreted visual evidence and sensitivity to textual formulation can produce a consistent but incorrect consensus that majority voting reinforces. Cross-view aggregation can expose such errors, but image and text transformations may themselves degrade task-relevant evidence to different extents. We introduce \method, a reliability-aware harmonic consensus framework for multimodal test-time reinforcement learning. \method constructs a $2\times2$ grid combining original and perturbed images with direct-solving and reframing modes, selects pseudo-labels through weighted harmonic aggregation, and adapts modality-level weights using batch-averaged, temporally smoothed shifts in pseudo-label confidence as a heuristic reliability signal. We formulate consensus as a weighted $χ^2$-barycenter problem whose unique minimizer is the normalized weighted harmonic
benchmark - arxiv:2610.39438 · cs.AIATLAS-Bench: An Agentic Tool-Use Benchmark for Local Area SearchKeyun Zhang, Dapeng Yin, Fanfu Wang, Shuang Lang +4
Reliable local search is essential for LLM agents to support everyday decisions and longer plans by finding places that meet users' needs at given times and locations. Yet existing benchmarks provide limited coverage of whether search results satisfy mandatory requirements, account for preferences, and ground factual and experiential judgments in place attributes and reviews. We introduce ATLAS-Bench, a benchmark of 1,200 queries across three regions, expanded from real search events using place attributes, reviews, and existing user profiles from the platform. Each query distinguishes mandatory requirements from preferences and factual conditions from subjective judgments, with supporting evidence and criteria for checking each requirement. Queries are validated through iterative assertion checks and cross-model review. We also introduce ATLASBox, a standardized evaluation environment with search tools and a fixed point-of-interest (POI) knowledge base. Given a query, timestamp, and l
agentllm agentagentictool-usebenchmark - arxiv:2610.39433 · cs.CLARBITRA: Robust Document-Level ATT&CK Extraction via Multi-Expert ArgumentationYoungjin Jin, Seungwon Shin
Cyber threat intelligence (CTI) reports describe adversary behavior in natural language, but converting them into structured MITRE ATT&CK mappings remains difficult: existing methods decompose reports to sentence-level granularity for extraction, while modern LLMs can work with full reports in a single pass but with noticeable performance degradation. We present ARBITRA, which orchestrates multiple role-specialized LLM experts that propose candidate ATT&CK entities from the full report, which are then verified through a per-candidate quantitative bipolar argumentation framework (QBAF) grounded in document evidence and rationales. We additionally construct reasoning-specialized corpora for fine-tuning so experts can act as effective participants during this verification process. Across ATT&CK extraction benchmarks, ARBITRA reaches a stronger precision-oriented operating point than sentence-level pipelines, recent LLM-based extraction systems, and frontier models including GPT-5.5. Under
benchmark - arxiv:2610.39429 · cs.ROScheduleGen: Demonstration Generation via Skill Scheduling for Multi-Arm Imitation LearningNeel Jawale, Stephan Pleines, Soha Pouya, Yuke Zhu +2
Multi-arm robot platforms, including bimanual and humanoid robots, have the capability to complete tasks with lower cycle times than single-arm systems, but they are more difficult to safely control due to the presence of multi-arm constraints. The paradigm of training generalist robot policies from massive datasets is a promising approach to address general-purpose multi-arm manipulation, but the acquisition of large multi-arm manipulation datasets remains a bottleneck. Automated demonstration generation can alleviate this burden by synthesizing new demonstrations from a handful of source demonstrations, but existing algorithms ignore multi-arm constraints and as a result often induce unsafe and unsuccessful behaviors. We address this limitation by introducing ScheduleGen, an integrated planning & scheduling algorithm for multi-arm demonstration generation. We evaluate ScheduleGen on 28 simulation task variants, including seven externally authored RoboDojo tasks and settings inaccessi
manipulationhumanoidsim-to-real - arxiv:2610.39419 · cs.CVDiagnosing the Reference Gap: What Visual Scaffolds Can and Cannot Fix in Multimodal ReasoningGuizhen Chen, Weiwen Xu, Wotao Yin, Anh Tuan Luu +1
Multimodal large language models can recognize visual content yet struggle to reason reliably over it across multiple steps. In counting, a model may see all objects but lose track of which ones it has counted. We argue that such failures, commonly attributed to perception, reveal a reference gap: the failure to bind language to the relevant visual entities, regions, or locations. Motivated by the theory of visual routines, we test whether external references such as numbering can narrow this gap. We introduce ReGap-Bench, a paired benchmark of five visual reasoning tasks in which the scene, question, and answer are fixed and only the reference varies. Results show that task-matched references improve accuracy by up to 25\%, while mismatched references provide little benefit or even hurt performance. These findings identify the reference gap as a bottleneck distinct from visual perception and show that constructing task-appropriate visual references can serve as a training-free visual
benchmark - arxiv:2610.39417 · cs.AIWell-Specified and Well-Tested? Assessing SWE-bench-Style Datasets with a Self-Evolving AgentZhonghao Jiang, Xuyang Zheng, Yanfu Yan, Zhongxin Liu +2
Large language models (LLMs) have shown strong performance on software engineering (SE) tasks, driving the need for high-quality datasets for reliable training and evaluation. SWE-bench-style datasets (SWE datasets) are widely used to evaluate LLM-based coding agents on real-world software issues, where task clarity and test coverage directly affect whether instances are solvable and verifiable. However, quality issues persist in existing SWE datasets, and manual annotation is costly and labor-intensive, motivating LLM-based annotation. Existing approaches typically rely on static, overly general rubrics, limiting their ability to handle diverse and nuanced scenarios and often yielding high false positive rates. In contrast, human annotators draw on accumulated experience to make context-aware judgments. Inspired by this, we propose ExLabel, an LLM-based, self-evolving, experience-guided assistant for task clarity and test coverage assessment in SWE datasets. ExLabel maintains an evolv
agentself-evolvingbenchmark - arxiv:2610.39416 · cs.RODecoding Geometry, Not Tokens: Efficient VLMs for Robot Affordance and Trajectory PredictionAkila Ayanthi, Darshana Priyasad, Tharindu Fernando, Dimity Miller +3
Predicting where to grasp and how to move is fundamentally a geometry problem. Vision-language models treat it as language generation instead. They tokenize spatial coordinates into text and decode them autoregressively. This mechanism ignores spatial structure and truncates continuous precision. Errors compound across coordinate tokens during decoding. In this paper, we show that compact off-the-shelf VLMs with 0.5-2B parameters already encode rich spatial information in their intermediate hidden states. Autoregressive decoding is the dominant bottleneck for trajectory prediction, while affordance prediction depends additionally on feature depth and decoder architecture. We expose the latent spatial information with a lightweight, geometry-compatible prediction head that attaches to a frozen backbone. It reads features from an intermediate hidden layer, conditions a learned query on the task instruction and refines it via cross-attention over the visual tokens to regress affordance bo
grasp - arxiv:2610.39412 · cs.LGDobb: Shard-Local Preconditioning in Block-Diagonal EigenbasesShi-Yuk Fong, Shuhua Yu, Jiawei Zhang
Training language models whose parameters and optimizer state exceed a single node's memory requires sharding these tensors across devices. Under tensor parallelism (TP) or fully sharded data parallelism (FSDP), matrix-aware optimizers such as Muon introduce communication to perform global matrix operations across shards. We propose Dobb (Distributed Optimizer with Block-diagonal Eigenbases), a matrix-aware optimizer designed to compute updates locally within parameter shards. Dobb approximates the input-activation second moment with diagonal blocks and tracks a complete basis within each block. It projects momentum into these bases, applies sign normalization to the coefficients, and reconstructs the update. When each block lies within one shard under TP or FSDP, Dobb applies its matrix update locally, requiring no communication beyond data-parallel synchronization of the activation second moments. We benchmark Dobb against Dion, a distributed matrix-aware optimizer designed to retain
memorybenchmark - arxiv:2610.39403 · cs.CVSymKL: A Unified Representation-Space Measure of Signal DegradationDhananjaya Jayasundara, Sudarshan Rajagopalan, Vishal M. Patel
Quantifying degradation between a pristine signal and its degraded observation is fundamental to signal processing, yet existing full-reference measures are inherently modality-specific. We introduce SymKL, a unified full-reference degradation measure based on implicit neural representations (INRs). An INR is fitted once to the pristine reference and characterized by its diagonal Fisher information, which captures the local information geometry of the learned representation. Evaluating the Fisher information with the pristine reference and a degraded observation yields two parameter-sensitivity profiles whose symmetric KL divergence defines the degradation score. The proposed formulation requires only the reference and degraded observation, without knowledge of the corruption process or degradation severity, making it applicable across images, audio, and 3D data. We further derive an exact closed-form expression for the diagonal Fisher of SIREN, enabling efficient computation. Experime
benchmark - arxiv:2610.39396 · cs.CVCASE: Concept-Aware Segmentation from Positive and Negative Visual ExemplarsYi-Jen Tsai, Yen-Yu Lin, Chien-Yao Wang
Few-shot semantic segmentation typically identifies target objects based on positive support examples of a single semantic class. In practice, however, positive cues alone are often underspecified: for instance, "red cars" may be confused with cars of all colors, "vehicles" may include cars that should be excluded, and "material-defined objects" may span multiple categories. Fulfilling precise requests requires positive exemplars to provide inclusion evidence and negative exemplars to act as intent-defining constraints for exclusion and disambiguation. We formulate this setting as Specified Visual Prompt Segmentation (SVPS), where the target concept is specified by both positive and negative visual prompts. In this work, visual exemplars refer to support-image regions specified by bounding-box prompts. To study SVPS, we introduce CASE (Concept-Aware Segmentation from Positive and Negative Visual Exemplars), a benchmark built on COCO and PACO. CASE encompasses six category- and attribut
benchmark - arxiv:2610.39395 · cs.AIOPERA: OPamp Design by an Expert-Taught Reasoning Agent, from Specifications to LayoutVishal Saxena, Soumyajit Mandal
Analog flows from specifications to a signed-off layout rarely run without manual iteration. Recent large language model (LLM) agents size circuits with optimizer loops over the device dimensions and stop at the schematic. We present OPERA (OPamp design by an Expert-taught Reasoning Agent), in which a general-purpose LLM agent designs CMOS opamps from specifications to post-layout sign-off with open-source tools. First, the designer teaches the agent the gm/ID method: map the specifications to transconductances and minimum currents, choose gm/ID per device role from characterized lookup tables, and size from equations, leaving a few knobs to search and an explainable design. Second, the agent learns the gLayout framework and taught layout practice, and writes a reusable layout generator that closes the layout and extraction loop at designer-approved gates. In SKY130, the agent designed a shared constant-gm bias and six opamps (5T, folded-cascode and telescopic OTAs; Miller, folded-casc
agentllm agentbenchmark - arxiv:2610.39385 · cs.ROUMI-Live for In-The-Wild Bimanual Robotics Data CollectionGovind Velamoor, Adrian Lin, Sam Jung, Spencer Mateega +2
Universal Manipulation Interface (UMI) devices allow human experts to demonstrate tasks for robot learning without teleoperating physical robots. Using these demonstrations as training data still requires a consistent hardware setup, calibration, and data processing. We present UMI-Live, a full-stack open-source bimanual UMI system for in-the-wild data collection. We design our custom UMI-Live hardware to integrate camera and LiDAR--inertial sensing, onboard compute, and power, without external tracking equipment or calibration within the collection environment. We implement a simple user interface for recording data while an asynchronous pipeline performs parallel SLAM reconstruction and publishes training-ready datasets. We also release software that supports dual-YAM teleoperation and VLA policy training and inference. We validate UMI-Live data quality by fine-tuning $π_{0.5}$ and MolmoAct2, achieving success rates of 10/10 and 8/10, respectively on a bimanual pick-and-place task. F
vlavla policymanipulationteleoperation - arxiv:2610.39380 · cs.AILearning to Formulate: Agentic Hardware-Software Co-Design of AI Infrastructure through Progressive Optimization and VerificationYi Ding, Li Shang
AI infrastructure, the chips and software that train and serve AI models, is designed in layers. Layers make the system manageable, but they also fix which choices an optimizer can consider. Gains that need the chip and its software to change together are excluded before optimization begins: the formulation leaves them out. Language-model agents make it possible to revisit these boundaries, but an agent's proposal cannot show by itself that a design is correct or optimal. We present Formulaic, a flow that divides the work. The agent writes the formulation, a mathematical solver optimizes within it, and verification and measurement check its assumptions and results. The agent refines the problem step by step, from the computation graph toward the chip. Each step is a template: a family of programs with the conditions under which they are correct. Wall relaxation searches the chip by pricing the hardware limits that bind the software. This paper takes a first step, through separate studi
agentagentic - arxiv:2610.39377 · cs.AIAgentDecept: Benchmarking Deceptive Behavior in Tool-Using LLM Agents under PressureYehan Zhang, Qianqi Yan, Hang Yin, Ching-Chen Kuo +4
LLM agents increasingly carry out multi-turn tasks across everyday applications, yet users often see only the agent's report rather than the observations, API calls, and state changes behind it. Existing deception benchmarks have studied long-horizon interactions and tool-use settings, but whether agents faithfully report their actions across realistic, stateful application workflows remains underexplored. To address this gap, we introduce AgentDecept, an executable benchmark for evaluating behavioral deception under pressure. AgentDecept contains 477 pressure variants derived from 257 source tasks across six pressure sources, each paired with a pressure-free counterpart. Interventions enter through event messages, task context, application state, tool responses, and service quotas, spanning explicit events and implicit task or environment changes. Our evidence-grounded protocol evaluates full trajectories by relating user-facing claims to time-ordered observations and executed outcome
llm agenttool-usebenchmark - arxiv:2610.39372 · cs.AICord: Confidence-Anchored Order Residual Decoding for Masked Diffusion Language ModelsChengwei Zhou, Zhaoyan Jia, Massoud Pedram, Gourav Datta
The decoding order of masked diffusion language models (dLLMs) can substantially affect generation quality, yet common decoders typically apply the same reveal rule throughout generation. Confidence provides a strong default, but it measures how decisively a position can be predicted now, not how useful revealing that position will be for subsequent decoding. We propose Cord, a lightweight prompt-conditioned policy that retains confidence decoding as a stable reference and learns when and how to depart from it through geometric residuals over generation direction, proximity to revealed context, and two-sided infilling. The residuals adapt to both the prompt and the evolving partial response. Rather than sampling a new policy action at every decoding step, Cord samples a single low-dimensional coefficient matrix per rollout, which induces an entire state-dependent reveal trajectory. This yields a contextual-bandit formulation trained directly from sequence-level rewards, without step-le
benchmark - arxiv:2610.39367 · cs.ROUtility-WAM: Correcting Future-Induced Action ChangesFeihong Zhang, Xi Lin, Yulong Shi, Zuxing Lu +6
World Action Models (WAMs) can condition robot actions on generated visual futures, but this additional context is neither cost-free nor uniformly beneficial. We examine what a predicted future actually changes: the action delta $Δa$ between the future-conditioned and future-free execution paths of the same model. Under a shared WAM checkpoint, this delta barely shifts average benchmark success ($+0.8$ points on LIBERO, $-0.8$ on RoboTwin~2.0), while generating visual futures increases inference latency by 1.7--1.9$\times$. These aggregate metrics conceal a concentrated failure mode that we term \emph{future-induced negative transfer}: while predictive deltas benefit long-horizon tasks ($+2.4$ points on LIBERO-Long) and remain negligible on most routine manipulations, they induce severe degradation on precision tasks---with \emph{Press Stapler} and \emph{Place Can Basket} losing 54.0 and 34.0 points, recurring as top failure modes across independently trained WAM variants. We introduce
manipulationliberorobotwinbenchmark - arxiv:2610.39358 · cs.AIBefore the Next Rollout: Diagnosing Skill Utility Shifts from Policy UpdatesYifan Wang, Zhenyi Gai, Yiwen Zhao, Mingxuan Jiang +4
Skill-evolution methods typically analyze execution trajectories to guide agent skill refinement. However, policy updates can change the behavioral effects and marginal utility of an unchanged skill. Attribution based on earlier trajectories may therefore no longer reflect the skill's contribution under the updated policy. Moreover, trajectory failures can arise from multiple factors, making individual skill contributions difficult to isolate. We evaluate whether policy-update signals contain predictive information about changes in skill marginal utility. At fixed contexts from pre-update training rollouts, we measure how policy updates change the contrast in token preferences between skill and control conditions. Combining these interaction shifts with advantage estimates yields reward-directed readouts for diagnosing utility changes. The readouts reuse existing training rollouts and require only forward passes for single-step action scoring, without additional post-update rollout eva
agent - arxiv:2610.39341 · cs.LGRetroChessBench: A Benchmark for Reconstructing and Proving Chess HistoriesTidor-Vlad Pricope
Frontier language models such as GPT 6 Astra have contributed to research mathematics, demonstrating substantial reasoning capabilities. Yet serious reasoning failures persist even in tasks governed by explicit rules. We demonstrate this in retrograde chess which is a playground for pure reasoning, where solvers reconstruct legal histories and distinguish possible events from forced ones. We introduce RetroChessBench, a 12-task benchmark covering exact-length constructions, unique endings, and exhaustive answer sets, with references, rubrics, and a terminal-reward environment. In two trials per task and condition, GPT 6 Astra High fails to produce the required final answer in 22/24 attempts without tools and 18/24 with code execution; mean terminal RL rewards are 14.75% and 27.19%, respectively. The examples identify faulty deductions: failed responses reject legal histories and treat temporary obstructions as permanent. Online retrograde archives often provide puzzles and final move s
benchmark - arxiv:2610.39339 · cs.CLDoes the Citation Matter? Quality, Grounding, and Physician-Rubric Accuracy of Google's AI Overviews for Medical InformationSam Agnoli, Brigid Reilly, Lily Barna, Emőke-Ágnes Horvát
Google's AI Overviews (AIOs) are the first result users encounter when searching the web, shifting the work of selecting and evaluating sources from the user to Gemini. This may be especially consequential for health information, given that Google's traditional algorithm was designed to prioritize quality sources, and inaccurate information carries health risks. We evaluated 3,611 benchmark health questions, comparing each AIO against that question's top five traditional results on source selection and quality. We also evaluated the AIOs' statement grounding in their cited sources and how accurate AIOs were against physician-written rubrics. Overviews appeared for 96.4% of questions but drew on substantially different evidence than traditional search: only 44.4% of the top five results were cited by the corresponding AIO, which instead overrepresented commercial health publishers (e.g., WebMD, Healthline) at the expense of academic and government sources (e.g., Cleveland Clinic, Nation
benchmark - arxiv:2610.39330 · cs.CVCan Image Agents Complete Visual Commissions? A Benchmark and Modular FrameworkTong Li, Lishuai Gao, Qi Zeng, Yifan Li +3
Visual commissions require image agents to integrate heterogeneous source materials, preserve requirements across dependent stages, and produce coordinated deliverables. Existing evaluations examine specific capabilities and execution settings, leaving end-to-end commission completion less understood. We introduce Visual Commission Bench (VCB), comprising 600 tasks across 12 categories and three complementary tiers, from capability-focused tasks to complete commissions. VCB uses 5,039 requirement checks and separate visual quality ratings to assess content correctness, visual realization, constraint preservation, and delivery completeness without prescribing an internal workflow. We present V-Protocol, a reference agent that dynamically composes reusable task, cognitive, and expression modules at inference time. Task modules organize delivery, cognitive modules derive content, and expression modules guide visual realization. The modules exchange structured intermediate artifacts, while
agentbenchmark - arxiv:2610.39327 · cs.ROSONIC-K1: Porting a Universal Humanoid Motion Tracker to a Small Humanoid on a Single GPUKyrylo Kolesnichenko, Jong-Hoon Kim
We present SONIC-K1, to our knowledge the first open-weights whole-body motion tracker for the Booster K1 humanoid, obtained by adapting the SONIC framework to this smaller robot. The policy is trained from scratch on a single shared GPU with a small fraction of the compute reported for SONIC. In MuJoCo simulation, it tracks nearly all held-out motion clips and performs well on external benchmarks under SONIC's success criterion. On the real K1, it tracks upper-body and slow motions, while stepping and turning remain less reliable, and simulation experiments show a sensitivity to action latency. These results show broad motion-tracking capability on a modest training budget and point to quantitative hardware evaluation as the main next step. Code, policy weights and evaluation tools are publicly available.
humanoidbenchmark - arxiv:2610.39324 · cs.CVMRC-MMAgent: Reinforcement Learning for Multi-Role Collaboration in Multimodal Tool-Using AgentsShengqin Wang, Jie Jin, Xiaofan Li, Wentao Yan +7
Multimodal tool-using agents augment vision-language models with access to external information and capabilities, enabling them to solve knowledge-intensive and visually complex tasks that cannot be addressed using visual context alone. However, existing reinforcement learning pipelines largely train single-role agents to search and answer using terminal rewards, overlooking how humans inspect retrieved evidence, reformulate queries, and revise conclusions during search. Consequently, they fail to optimize this self-corrective process over complete interaction trajectories. We propose MRC-MMAgent, a multi-role collaborative reinforcement learning framework that incorporates trajectory assessment and continuation directly into the executable agent loop. First, the framework establishes a multi-role collaborative loop that integrates solving, inspection, and continuation, enabling the agent to acquire evidence through tools, evaluate the preceding trajectory, and resume problem solving w
agentbenchmark - arxiv:2610.39323 · cs.LGClass-aware Latent Attention Replay (CLAR): Class Incremental Hazard Diagnosis for Resilient AI in the Manufacturing Industrial InternetAvranshu Mahanty, Ran Jin, Yingyan Zeng
Artificial Intelligence (AI) systems provide critical computation services in the Manufacturing Industrial Internet (MII), and their resilience depends on timely diagnosis of hazards across the data, AI pipeline, and cyber-physical layers. During operation, new hazard types keep emerging with only a few samples each, hazards co-occur, and the diagnosis model runs on memory-constrained computation nodes. Replay-based class-incremental learning (CIL) can learn new hazard classes while revisiting a small memory of past samples, but existing methods assume single-label, balanced streams, store high-dimensional raw samples, and select them independently of the classifier. We propose Class-aware Latent Attention Replay (CLAR), which combines a dual buffer that stores mostly compact latent representations with a few raw samples, an Attention Leverage Score (ALS) that upper-bounds the classifier logit and guides class-balanced reservoir sampling, and an exponential-moving-average teacher that
memory - arxiv:2610.39322 · cs.CVMemoryFusion: Cross-Temporal Memory Learning for Multimodal Video FusionMeiqi Gong, Hao Zhang, Jiayi Ma
Multimodal videos contain rich temporal cues that can facilitate faithful fusion. However, existing video fusion methods mainly rely on short-range temporal interactions, leaving valuable long-term information underexplored. Moreover, inconsistent fusion representations can cause flickering artifacts, undermining temporal coherence. To address these coupled challenges, we propose MemoryFusion, a two-stage framework integrating cross-temporal memory learning with progressive temporal refinement. Considering the complementary roles of adjacent and distant frames, the first stage combines a Short-Term Memory Bank for local temporal guidance with a Long-Term Memory Bank that preserves reliable historical information through a fusion-oriented update strategy. To mitigate temporal fluctuations, a Temporal Smoothing Module leverages soft motion cues for coarse background stabilization. Building upon the resulting fusion prior, The second stage further employs a lightweight 3D residual Tempora
memory - arxiv:2610.39320 · cs.LGBEVLight: Generalizable Traffic Signal Control from UAV Bird's-Eye ViewsMaonan Wang, Yuxin Cai, Junming Liu, Zhengyan Huang +4
Learning-based traffic signal control (TSC) has achieved promising performance, but most methods rely on structured traffic states that must be estimated from sensor observations in real deployments. Visual TSC learns signal decisions directly from images, yet existing methods depend on fixed roadside cameras and remain tied to specific intersection configurations and visual domains. UAVs offer on-demand bird's-eye-view (BEV) observations that jointly capture intersection topology and traffic dynamics, making them a promising sensing platform for visual TSC. We present BEVLight, a generalizable TSC framework that directly maps BEV observations to signal actions. BEVLight grounds visual features onto individual lanes and composes them along the junction topology into movement and phase representations, allowing a shared policy to score candidate phases across heterogeneous intersection layouts and signal plans. To train the policy, we develop a web-based virtual teleoperation platform a
teleoperation - arxiv:2610.39314 · cs.ROLanguage-Grounded Force/Torque Pretraining for Extrinsic Contact PerceptionKazuki Takahashi, Hiroaki Murakami, Ryota Sakuma, Mitsuhiro Kamezaki +2
Fine manipulation requires robots to recognize when a grasped object makes or changes contact with its environment, yet such extrinsic contact is not always directly observable from vision or local contact geometry. Force and torque signals carry this information, but also vary with object weight, grasp configuration, and robot motion, leaving it unclear whether they can provide a reusable representation of physical state. Here we show that bilateral six-axis force/torque histories can be pretrained as a semantically structured physical representation by aligning them with a frozen language space using supervision generated from simulated pick-and-place interactions. On unseen objects in simulation, the learned representation supports text-prototype readout of robot action, object weight, and grasp interaction. In real-world placement, unlabeled calibration enables contact and support-state recognition without updating the encoder, outperforming a vertical-force sign-reversal rule base
manipulationgrasp - arxiv:2610.39308 · cs.CVUnlimited OCR Works 2Shu Wei, Jingjing Wu, Qunyi Xie, Lingshu Zhang +8
Unlimited OCR ushers the era of long-horizon parsing, substantially mitigating the growth of working memory as document lengthens. However, as document length scales to hundreds of pages, the visual context grows substantially, sharply increasing the computational cost of each decoding step and limiting the scalability of ultra-long document parsing. Inspired by human book transcription via localized focus without re-scanning preceding pages, we introduce Unlimited OCR 2, a document parsing model that looks only at what it transcribes. Building on Unlimited OCR, we upgrade all the decoder attention modules to Sparse Reference Sliding Window Attention (SR-SWA), bounding active reference tokens to a fixed capacity and thus sustaining constant computational costs across arbitrarily long documents. By combining SR-SWA with Unlimited-Horizon Decoding Paradigm, Unlimited OCR 2 enables document parsing over virtually unlimited lengths without prohibitive memory overhead. At the hundred-page s
memorybenchmark - arxiv:2610.39305 · cs.LGLearning to Bound: Monotonic, Calibrated Neural Upper Bounds for MPE InferenceAkshay Vyas, Hailiang Dong, Shivvrat Arya, Brij Malhotra +3
Probabilistic graphical models (PGMs) compactly encode structured joint distributions, but Most Probable Explanation (MPE), which finds the highest-probability assignment to unobserved variables, remains computationally intractable in large, densely structured models. Anytime systematic search, such as branch-and-bound (B&B), is a popular approach, but its efficiency depends on tight admissible upper bounds that are often prohibitively expensive to compute using classical methods like Mini-Bucket Elimination (MBE). We introduce Learning to Bound (L2B), a data-driven framework that amortizes bound computation cost by training a neural predictor. L2B targets two key properties required for effective search bounds: admissibility, which is achieved via statistical calibration techniques, and progressive tightening as evidence accumulates, which is enforced through structural architectural constraints that ensure monotonicity by construction. Across challenging benchmarks, L2B delivers tigh
benchmark - arxiv:2610.39302 · cs.CVProving and Improving the Platonic Representation Hypothesis with JEPAsRandall Balestriero, Yann LeCun
Whether models trained on different modalities converge to the same representation of the world is central to AI. The Platonic Representation Hypothesis (PRH) conjectures that they do as they scale, but without a proof it offers neither a guarantee nor a recipe. We turn the hypothesis into a theorem. First, we prove a \emph{Platonic Representation Theorem}: Joint-Embedding Predictive Architectures (JEPAs) trained separately on any number of modalities learn the same representation up to one rotation whenever their positive pairs are two observations of one scene--\emph{good views}--and their misalignment is bounded otherwise. Second, we show that next-token prediction is a JEPA with \emph{bad views}, parts of a single document, which explains why language models align with vision only through scale. Third, we validate the theory from 145k captions to the texts of 940M web images, and uncover a scaling law of alignment that predicts unseen runs within seed noise. Good views need 42 time
benchmark - arxiv:2610.39301 · cs.CVYoutu-Parsing-Omni: One Encoder, One Schema, All ModalitiesYunhang Shen, Zhangquan Chen, Xiaotian Li, Zhongpeng Cai +26
We present Youtu-Parsing-Omni, an omni-modal parser that converts documents, images, audio, and video into structured records grounded in space and time, providing a perception layer for agents that read, locate, and cite evidence across modalities. Its design rests on one encoder for every modality and one schema for every task. Youtu-Omni-Encoder, initialized from a pre-trained language model to strengthen text reading, encodes all modalities with shared weights in a unified space-time coordinate system, and a few fusion layers let video frames attend to concurrent audio, bringing cross-modal fusion and temporal alignment into perception itself. OmniSchema unifies the outputs of all ten parsing tasks, allowing a single self-evolving post-training procedure, OmniSchema-Aware On-Policy Distillation (OSAD), to provide a shared token-role-based objective across tasks. In each round, the model is trained on its own rollouts against token-level targets from a reference-conditioned copy of
self-evolvingpost-training - arxiv:2610.39292 · cs.LGStrongly Adaptive Online Learning Achieves Uniform Discounted RegretSifan Yang, Wenhao Yang, Lijun Zhang
Discounted regret measures the performance of an online learner by assigning exponentially decaying weights to past losses, reflecting the greater importance of recent losses in non-stationary online learning. While previous work obtains uniform discounted regret bounds over a continuous range of discount factors using a specialized meta-algorithm, its results are limited to convex functions. In this paper, we demonstrate that strongly adaptive regret bounds can directly imply uniform discounted regret guarantees for convex, strongly convex, and exp-concave loss functions. Our key observation is that discounted regret against a fixed comparator admits an exact decomposition into a convex combination of interval regrets. For convex functions, this reduction yields an $O(\sqrt{\log T/(1-λ)})$ discounted regret bound for all discount factors $λ\in(0,1)$, matching the previous rate over a broader range of discount factors. For strongly convex and exp-concave functions, we provide regret bo
online learning - arxiv:2610.39287 · cs.ROOpenFlyV2: An onboard vision-language-action model for aerial perception and navigationJunli Liu, Zhongrui You, Chenhui Li, Zhen Li +8
Human cognition flexibly draws on a shared perceptual foundation to understand environments, attend to referenced entities, reason about spatial relationships, and act. Endowing aerial robots with an analogous functional organization remains challenging, as existing systems are often specialized for individual capabilities, dependent on offboard computation, or unable to sustain closed-loop operation on resource-constrained platforms. We present OpenFlyV2, a unified onboard vision-language-action architecture that uses a shared vision-language representation to support semantic understanding, language-grounded perception, spatial reasoning, and continuous aerial control within a single model. A compact action expert converts the shared multimodal context into short-horizon flight waypoints, while motion-aware temporal memory and action priors reduce the computational cost of closed-loop execution. We construct \dataset{}, a simulation-based dataset comprising approximately 70,000 navig
vision-language-actionmemory - arxiv:2610.39282 · cs.AIWhat Does "Now" Mean to Your Computer-Use Agent? Stale Screenshots, Page Loads, and Which Check to Run Before a ClickVasundra Srinivasan
Computer-use agents plan on a screenshot for several seconds and then click at a position on the screen; if the page has changed in the meantime, the click lands on an unintended element. Prior work treats the planning interval as the source of this risk. Measuring planned clicks on 31 public web pages, we find that the risk lies in page loading: 0.4% of clicks planned on fully loaded pages became stale within 30 seconds, against 17% of clicks planned in the first second after a page opened. On these pages, waiting about one second for network activity to stop removes most of this risk, yet the wait rules of eight of ten widely used open-source agents leave more than 1% of clicks stale. Commercial pages keep loading for longer: in the HTTP Archive's public crawl, 15% of 65 desktop booking and buying home pages went 500 ms without network requests within 3 seconds of their HTML being parsed, against 63% of 3,028 .gov pages, and 52% shifted their layout within 6.5 seconds of a network-id
agentbenchmark - arxiv:2610.39279 · cs.ROCuBot-STT: Simultaneously Transport and Tighten for Lattice-type Modular Reconfiguration RobotChongxi Meng, Da Zhao, Yanmin Zhou, Runjie Shen +2
Lattice-type modular reconfigurable robots rearrange standardized modules across tasks. Dedicated transport robots reduce module weight and cost by eliminating module level reconfiguration actuators. But existing systems use low capacity connections or internal fastening robots that occupy space for functional components. This paper presents CuBot-STT, which integrates module transport, interface locking, and release in one external robot. Cubic modules provide structural support and $24\,\mathrm{V}$ power distribution. On each face, a gear transmission converts one external input into motion of four helical locking points, preserving internal space. The robot uses alternating foot support, a gripper for module handling, and a tightening head for interface actuation. The controller coordinates two-foot support and module locking through continuous constraint transfers. The planner decomposes relocation costs into an analytical lower bound and a residual learned through expert-guided, g
manipulationgripperbenchmark - arxiv:2610.39278 · cs.ROWorldFlow: Joint Video and 3D Motion Prediction via Large-Scale PretrainingYuhao Lu, Jiongxu Chen, Hexiang Wei, Jiajin Wu +8
Predicting future physical states stands as a central challenge in embodied intelligence. Existing video world models leave 3D dynamics implicit, action-based formulations remain tied to specific robot configurations, and 3D motion prediction models use flow-only supervision, providing a limited learning signal for physical dynamics. To circumvent these limitations, we introduce \method{}, a generative world model that jointly predicts future video and query-based 3D trajectories. By parameterizing 3D motion as continuous displacement tokens aligned with the video grid, a unified Mixture-of-Transformers (MoT) backbone jointly denoises both modalities under a rectified-flow objective. This formulation couples explicit geometric dynamics with visual evolution, effectively unlocking the physical commonsense embedded within pretrained video models. To empirically scale and validate this unified architecture, we train \method{} on a heterogeneous mixture of 1.44 million object-level samples
embodiedworld model - arxiv:2610.39276 · cs.AIMemSteer: Turning User Histories into Supervision for Agent Memory OptimizationJie Huang, Cuiling Lan, Qize Yu, Xiulian Peng +1
An effective agent memory system should adapt to a user's history and task requirements. Human-annotated supervision can guide this adaptation by providing feedback for evaluating and improving memory designs. Yet suitable labels are rarely available for new user histories, and manually constructing QA pairs and supporting evidence is costly. We introduce MemSteer, which turns unlabeled user histories into question-answer (QA) supervision with supporting evidence for memory-design optimization. MemSteer constructs an evidence-grounded answer plan that specifies how facts across related records jointly support a reference answer. The plan guides question generation, while source-grounded auditing and blind answer recovery check the resulting annotations. To study the utility of this supervision, we develop MemSteer-Opt, which combines QA-based design evaluation with evidence-guided failure diagnosis to revise memory store, retrieval, and context construction without updating model param
memoryagent memoryagent - arxiv:2610.39267 · cs.RORAIBO-Wagon: Real-Time Navigation System for a Legged Robot Towing a WagonInsung Yang, Eunchae Kim, Jungwoo Hur, Seunghun Jeon +2
Legged robots have demonstrated advanced capabilities across a wide range of real-world tasks, yet much of this progress concerns the robot controlling only itself. Once the robot is coupled to an external object -- particularly an underactuated one -- the dynamics of the combined assembly become more complex. Towing a wagon is a common practical scenario, where coupled dynamics make reliable navigation challenging. This difficulty is amplified in confined spaces, where collision-free navigation requires finding a narrow set of feasible paths under strict real-time constraints. To address this, we propose a navigation framework that combines Conditional Flow Matching (CFM) with Model Predictive Path Integral (MPPI) control over a learned Forward Dynamics Model (FDM). The CFM model mitigates the initialization problem of MPPI and accelerates its convergence, while MPPI with the FDM recovers feasible solutions beyond the CFM model's training distribution by iterative refinement. We integ
quadrupediterative refinement - arxiv:2610.39265 · cs.CVDeCoRL: Decomposed Consistency Reinforcement Learning for Multi-Image CompositionZhiqiang Wu, Shuang Sun, Jiale Zhang, Jing Li +3
Multi-image composition (MICo) is a critical yet challenging capability for large-scale image editing models, where consistency degrades rapidly as the number of reference images increases. While supervised fine-tuning (SFT) can equip models with basic MICo ability, it provides only implicit supervision on the composed image and often fails to learn reliable object-level correspondence across multiple references. We propose DeCoRL, a reinforcement-learning (RL) framework that improves MICo consistency by decomposing holistic MICo consistency into object-level consistency. DeCoRL first establishes instruction-relevant object correspondences across reference images and localizes the corresponding regions in the generated image. It then uses a VLM-based reward model trained on single-object consistency annotations to produce fine-grained, object-wise feedback, which is aggregated into a multi-object consistency reward to guide RL post-training. We further introduce DeCo-Bench, a benchmark
post-trainingbenchmark - arxiv:2610.39264 · cs.AIOnline Dense Reinforcement Learning with Planning-Grounding Agent for UI Navigation on Live IOS AppsPeitong Duan, Khashayar Gatmiry, Jeffrey Nichols, Samy Bengio +1
While frontier vision-language models demonstrate strong capabilities in user interface (UI) navigation, their high inference costs prohibit everyday use. Conversely, smaller models trained via supervised fine-tuning (SFT) or offline reinforcement learning often memorize static trajectories, leading to brittle execution and an inability to recover from errors. To address these limitations, we propose a novel online reinforcement learning framework that trains UI navigation agents directly on live iOS applications. Our framework decouples the agent into two specialized components: a Planning agent and a UI Grounding model. We rely on an LLM-as-a-Judge to provide comprehensive reward signals, combining dense, step-level feedback to evaluate intermediate progress with a terminal reward for overall task completion. To effectively leverage this dual reward structure, we optimize the agent using Proximal Policy Optimization (PPO) with step-level advantage estimation. Furthermore, we introduc
agentbenchmark - arxiv:2610.39247 · cs.AISocialArena: Evaluating Social Intelligence in Large Language ModelsZongqi Wang, Rui Wang, Yuchuan Wu, Hua Yang +6
We introduce SocialArena, a benchmark for systematically defining and evaluating the social intelligence of large language models (LLMs). Although LLMs are increasingly deployed in socially interactive applications such as mental health counseling, companionship chatbots, and educational tutoring, existing evaluations remain fragmented focusing on isolated capabilities or single scenarios and evaluation reliability is largely unverified. This work makes three main contributions to address these limitations. (1) We first propose a unified three-tier framework that defines social intelligence along three capability dimensions: role-playing ability, mental state reasoning, and behavioral reasoning. To operationalize this definition, we design a role-playing dialogue paradigm with explicit character thinking for evaluation, enabling simultaneous assessment of these capabilities in a manner that mirrors realistic human interactions. (2) To ensure diversity across social scenarios and thereb
benchmark - arxiv:2610.39246 · cs.LGIMRouter: Imbalance-aware Temporal Pattern Routing for LLM-based Time Series ForecastingXin Yang, Cai Xu, Haishun Chen, Ying Liu +3
Accurate time series forecasting (TSF) is essential for Web applications, such as traffic monitoring and online resource management. Recent advances in leveraging pre-trained large language models (LLMs) have shown promising performance in TSF. However, existing LLM-based TSF methods largely optimize the overall forecasting objective while overlooking temporal pattern imbalance (TPI), where different temporal patterns occur at highly unequal frequencies. Our empirical analysis reveals that TPI is prevalent across widely used TSF benchmarks, with substantially higher forecasting errors on less frequent patterns. To address this issue, we propose IMRouter (Imbalance-aware Temporal Pattern Routing), a novel framework that mitigates TPI-induced biases in both temporal representation learning and cross-modal alignment. Specifically, IMRouter introduces Pattern-Adaptive Temporal Representation Learning (PATR), which leverages gradient conflict and local gradient density to guide expert speci
benchmark - arxiv:2610.39237 · cs.CLGrounding Multimodal Reasoning with Evidence-Ablated NegativesChenxi Liu, Tianyi Xiong, Yanshuo Chen, Heng Huang
Reinforcement learning with verifiable rewards (RLVR) has become a powerful tool for improving multimodal reasoning. However, standard outcome supervision suffers from credit assignment ambiguity, as a model may receive positive rewards by exploiting language priors or dataset shortcuts rather than grounding its reasoning in the requisite visual evidence. This makes high-quality negative responses critical, since effective negatives should remain plausible under the current policy while exposing the shortcut that should be suppressed. In this paper, we propose EVAN, a multimodal reinforcement learning framework utilizing EVidence-Ablated Negatives to train LMMs against their own language-prior failures. EVAN first identifies highly visually dependent examples using a scaled image information gain metric and preserves data diversity with a mixed-ranking strategy. It then constructs evidence-ablated images by masking the most discriminative image patches, eliciting plausible but visually
benchmark - arxiv:2610.39234 · cs.CLAgentHeal: Evaluating Goal-Preserving Resource Recovery of Web Agents under Access BarriersZhiyi Liu, Haohua Du, Haoqi Sun, Haisheng Tan
Large language model (LLM) agents have demonstrated strong capabilities in automating complex information-seeking and transactional tasks across the open web. However, their real-world reliability remains limited by access barriers, such as dead links, authentication walls, rate limits, and anti-scraping defenses. These barriers block the primary execution path, causing agents to abort prematurely, loop indefinitely, or quietly compromise the user's original constraints. Existing evaluations either implicitly assume that target resources are freely accessible or focus on transient perturbations that simple retries can resolve, failing to capture scenarios where agents must actively reroute around impassable bottlenecks. To bridge this gap, this paper introduces AgentHeal, a comprehensive benchmark evaluating goal-preserving resource recovery, defined as the ability of web agents to detect non-transient access barriers, pivot to legitimate alternative sources, and fulfill the user's exa
benchmark - arxiv:2610.39232 · cs.ROSeeing and Imagining as One: Unified Video Generation and 4D Perception for Embodied World ModelingHaitang Feng, Zhuoman Liu, Zhennan Wang, Sheng Wang +4
Embodied world models should not only imagine plausible futures, but also perceive the geometric, semantic, and motion structure of the evolving world. We present GenSight, a unified embodied world model that brings future video generation and dense perception into a single pretrained video diffusion transformer. GenSight uses a unified spatiotemporal representation for video, depth, embodiment segmentation, and first-frame-relative 2D flow. Built on a shared DiT backbone, GenSight uses lightweight task modulation to predict all modalities within the same architecture. All tasks are jointly optimized within a unified velocity prediction space. At inference, GenSight supports two complementary modes: multi-step future video generation with aligned dense predictions and single-step dense perception from an observed video. By unifying modality representation, model architecture, and prediction space, GenSight avoids separate modality-specific pipelines and couples appearance, geometry, se
embodiedmanipulationworld model - arxiv:2610.39230 · cs.AIBeyond LLM Wrappers: Context Engineering and the Three-Engine Architecture for Reliable Financial AIPradeep Kumar GS, Ravi Tanniru
Financial AI systems must support decision-bearing questions involving payments, collections, cash flow, and compliance while ensuring numerical accuracy, traceability, and auditability. Large language models (LLMs) enable natural language interaction but cannot independently guarantee reproducible calculations or evidence-grounded recommendations. We present a production-grounded architecture integrating three components: a financial knowledge graph for enterprise relationships, temporal context, and provenance; a deterministic optimization engine for financial calculations and decisions; and a multi-agent LLM layer for intent interpretation, task orchestration, explanation, and claim validation. A coordination layer, termed context engineering, selects, ranks, composes, and validates query-specific context across ERP systems, tools, and tenants. Based on observations from 15 deployments over approximately 11 months, single-agent pipelines exhibited hallucination rates of about 7-9% (
retrieval-augmentedknowledge graphmulti-agent - arxiv:2610.39212 · cs.ROEvent2Reward: Learning Robotic Rewards from Progress RelationsJunnan Nie, Zitian Fan, Jiachen Zhang, Junyi Lao +13
Learning task progress from robot demonstrations provides a practical way to construct dense shaping rewards for robotic reinforcement learning. A common approach uses normalized demonstration time as progress labels, implicitly assuming monotonically increasing task progress. During policy exploration, however, actual task progress can stall or even regress. Such models may mistake regressions for progress, assigning positive shaping rewards to actions that cause progress loss. Real suboptimal and failed executions offer corrective experience, but assigning absolute progress values to their states is difficult. Our key observation is that stalls, regressions, and recoveries establish local relative progress relations even when absolute progress is unknown. We introduce Event2Reward, which uses normalized temporal labels to anchor the progress scale and these within-trajectory relations to constrain local changes, combining absolute and relative progress supervision. Its progress predi
manipulationlibero - arxiv:2610.39205 · cs.LGSparse Autoencoders as Implicit Diversification Regularisers for Cross-Sectional Cryptocurrency ReturnsWeijia Han, Lisha Qu
Factor models are central to asset pricing and portfolio construction, but neural factors are often difficult to interpret, and their effect on portfolio risk is unclear. A sparse autoencoder (SAE) may address both concerns by describing daily returns across many assets with only a few active features. We ask whether this sparse representation produces stable financial factors and better-diversified portfolios. We study cryptocurrency returns, use equities as a secondary test, and compare the SAE with common benchmarks in a time-ordered evaluation. We find consistently lower volatility in SAE portfolios and link this result to lower correlations among the selected assets. However, the SAE does not outperform a simple contrarian strategy on risk-adjusted return, and its factor identities are not stable across training runs. Our contribution is to show that sparse coding can support portfolio diversification, while clearly separating this risk effect from claims about alpha or stable fac
benchmark - arxiv:2610.39201 · cs.CVCOGS: A Diagnostic Benchmark for Attribute-Aware Object-Level 3D Gaussian Splatting PerceptionZheng Xiao, Yifan Liu, Weijian Deng, Jiashu Li +3
3D Gaussian Splatting (3DGS) is increasingly explored as an input representation for 3D perception. Yet it remains unclear how current 3D architectures exploit heterogeneous Gaussian attributes. This ambiguity makes accuracy hard to interpret and limits principled architecture design for 3DGS perception. However, existing object-level 3DGS datasets lack the appearance variation or artifact control needed for systematic attribute-level analysis. We introduce COGS, a large-scale 3DGS dataset with 15,675 artifact-filtered and appearance-rich objects across 40 categories. Built on COGS, we establish COGS-Bench for object classification and attribute-usage diagnosis, complemented by external validation. Our benchmark shows that Gaussian attributes are not uniformly beneficial: selected subsets can outperform the full attribute set, and the effects of appearance, opacity, and covariance-related attribute combinations vary substantially across architectures. We further study this finding with
benchmark - arxiv:2610.39194 · cs.LGZero-Complement Extrapolation: Parameter Recovery and Classification in Severely Imbalanced Logistic Regression under Rank-Deficient RegimesChongsun Park
Parameter estimation and classification in binary logistic regression deteriorate severely when the number of minority class observations ($n_1$) is smaller than the feature dimension ($p$). Under such rank-deficient rarity ($n_1 < p$), maximum likelihood estimators suffer from quasi-complete separation, penalized likelihood methods exhibit severe intercept bias, and convex minority oversampling techniques such as SMOTE collapse into an $(n_1 - 1)$-dimensional affine subspace. In this paper, we propose a novel \textit{Self-Calibrated Zero-Complement} estimator that shifts the paradigm from minority interpolation to majority fringe extrapolation. Grounded in Stein's Lemma, we prove that the conditional expectation shift of the abundant majority class ($Y=0$) induced by logistic truncation uniquely identifies the true hyperplane normal $\boldsymbolβ$ after Mahalanobis whitening. We further establish that extrapolating synthetic minority instances outward from the high-quantile density fr
benchmark - arxiv:2610.39193 · cs.AIGoal-Aware Dynamic Multi-Agent Collaboration for Long-Horizon TasksXu Shen, Lingyong Yan, Shuaiqiang Wang, Jianmin Wu +3
LLM-based agents are increasingly used to complete complex, long-horizon tasks without human intervention. To make progress toward the task goal, agents need to perform different duties, such as planning and verification, when needed. To understand how single agents manage these duties during long-horizon execution, we conduct a preliminary study. Our analysis shows that verification has the strongest association with task success, but often occurs late, leaving limited opportunities for correction. We also find that the proportion of duty-relevant information decreases as records from different duties accumulate in a shared context. These findings lead us to examine whether multi-agent systems (MAS) can address these problems by assigning duties to agents with separate contexts. Our analysis shows that this separation alone is insufficient. Agents can lose track of the overall goal and repeat prior investigations. They may also lack necessary background while receiving irrelevant info
multi-agentagent system - arxiv:2610.39183 · cs.CVMulti-level Constrained Learning for Underwater Image EnhancementWangtao Bao, Liping Xie, Kanjian Zhang, Huimin Lu
Underwater image enhancement aims to correct color and recover detail degraded by wavelength-dependent attenuation, scattering, and noise. Paired-reference supervision constrains output similarity, but does not by itself specify the color mapping, the strength of local detail corrections, or consistency with the degraded observation. We propose MCL-UIE, a paired-supervised framework that addresses these requirements by constraining restoration operations and the training objective. Monotonic Chromatic Transform (MCT) restricts global color mapping in log-LMS space through positive diagonal entries and bounded cross-channel coefficients. Uncertainty-Gated Wavelet Reconstruction (UGWR) modulates high-frequency residuals with a learned confidence gate calibrated against normalized wavelet reconstruction errors. During training, Reference-Anchored Degradation Replay (RADR) renders the enhanced image and its paired reference using shared latent degradation proxies, encouraging both to repro
benchmark - arxiv:2610.39182 · cs.AIReasoning or Pattern Exploitation? Mechanistic Insights from RL-Trained Language ModelsAishik Nagar, Cheston Tan
Large language models (LLMs) are increasingly described as acquiring "reasoning" skills after reinforcement learning with verifiable rewards. However, whether these gains reflect genuine reasoning or more superficial adaptations remains underexplored. In this work, we probe LLMs trained in a finite and exhaustively analyzable logical domain, namely Tic-Tac-Toe, and trace how internal representations evolve across reinforcement learning (RL) with Group Relative Policy Optimization (GRPO). Quantitatively, RL yields high accuracy and robustness across prompt variations, suggesting the emergence of game reasoning. Mechanistic interpretability, however, paints a different picture: features extracted with sparse autoencoders (SAEs) reveal that models primarily adapt to rely more heavily on information already explicit in the prompt, such as whose turn it is, game progression and board occupancy. In comparison, high-level concepts like board symmetries and strategic forks remain weakly repres
post-trainingbenchmark - arxiv:2610.39180 · cs.ROEmbodiExp: Compositional Experience Transfer for Embodied AgentsZhifeng Gu, Ying Chen, Rui Jiang, Lei Yi +1
Experience can help embodied agents turn high-level plans into grounded actions, but transferring it across tasks remains challenging: previously effective strategies may be inapplicable to the current state or fail to produce their intended outcomes. We present EmbodiExp, a framework for constructing transferable experiences and composing them online during embodied task execution. Our key observation is that tasks often share subgoals whose completion can be assessed independently of overall task success. We therefore introduce an outcome-grounded subgoal experience that specifies a subgoal, applicability conditions, an adaptable policy, and a verifiable outcome. During execution, applicable experiences guide action generation, while verified outcomes can establish the conditions required by subsequent experiences, enabling online composition. Subgoal-level outcome attribution further enables failure utilization: failed trajectories can still contribute verified local successes and c
embodiedembodied agentbenchmark - arxiv:2610.39167 · cs.ROEgoContact: Evaluating Physical Intelligence with Egocentric VideoAdam Deryło, Alexey Gavryushin, Daniel Cremers, Marc Pollefeys +1
Egocentric video has emerged as one of the most scalable sources of data for learning generalizable manipulation priors. Yet, despite its growing use in pretraining robotic policies, it has been underutilized as a validation tool for physical intelligence, a key component of which is understanding where to make contact with an object to achieve an intended interaction. We present EgoContact, a large-scale egocentric dataset for evaluating the generalizability and precision of learned manipulation priors through future contact-point prediction. EgoContact contains 22.5K manually verified interaction events, each with a pre-contact observation, semantic intent, and pixel-precise contact locations. We find that current methods often struggle to predict precise interaction contacts and sometimes fail to localize the target object in diverse, in-the-wild scenes, even when allowed multiple predictions. Furthermore, we demonstrate the usefulness of EgoContact's annotated interactions for robo
manipulationgraspmemory - arxiv:2610.39162 · cs.LGA Novel Approach to Evaluating Evaluation Metrics for Multi-Output Structured PredictionAkshay Vyas, Angelo Pimienta, Nicholas Ruozzi
In multi-output structured prediction tasks, while only one ground truth label may be provided in the training data, multiple equally valid outputs may be possible, making reliable evaluation a persistent challenge. We postulate that human evaluators implicitly use task-specific invariants, e.g., object boundaries in colorized images or named entities in translations, to judge if an output is acceptable. Under this assumption, we introduce a notion of approximate task-specific invariants and use them as diagnostic tools to evaluate a variety of existing metrics for vision and language tasks. We use these task invariants as part of a framework to systematically test metric reliability by encouraging domain-relevant invariants in model outputs via an augmented loss function. In our experiments, we observe that enforcing invariants with an augmented loss yields substantial improvements in popular distributional metrics while more traditional metrics change only marginally. Through this in
evaluator - arxiv:2610.39158 · cs.CVVEvo-Harness: Causally-Guarded Self-Evolving Harness for Agentic Image GenerationYuan Wang, Yuhang Yu, Mengting Chen, Tianrun Xu +3
The central bottleneck in visual generation and editing has migrated from achieving pixel-level fidelity to satisfying densely constrained user specifications through increasingly agentic workflows. Prevailing agentic generation pipelines remain shallow generate-verify-rewrite loops treating VLM judgments as coarse visual feedback, either discarding repair experience or admitting ungated updates that poison skill library. With memorize and govern periods left open, they degrade into ungoverned iterators of repeated trial-and-error rather than continual self-evolving harness. Towards this issue, we present \textbf{VEvo-Harness}, the first causally-guarded continual self-evolving harness system for agentic visual generation. VEvo-Harness introduces four coupled innovations: (1) \textit{a visual traceback} that densifies sparse judgments into root-cause attributions with a hierarchical verification tree; (2) \textit{a visual continual governance} that admits only causally valid skill patc
agenticself-evolving - arxiv:2610.39155 · cs.AIREALM: Identifying Minimal Task Assets for LLM AgentsZihou Ren, Byungchul Tak, Long Wang
Large language model (LLM) agents often receive access to more files, accounts, and services than a task requires. Prompt injection turns this excess access into an opportunity to redirect execution toward unrelated resources. Applying least privilege therefore requires identifying the assets needed by a task, including assets whose identities emerge only during execution. We present REALM (Required Execution Assets Linked to a Mission), an approach to identifying and maintaining a task's required assets and their permitted uses. The approach represents required assets as Dependency Graph nodes, resolves prerequisite relationships, and admits newly discovered assets through source-scoped selection. It separates discovering an asset from authorizing its use, while preserving the agent's freedom to choose its computation. The REALM Builder maintains the selected grants and projects Asset Files; the Partial API Manager exposes resource-bound Partial APIs to the Agent inside a Task Contain
agentllm agent - arxiv:2610.39151 · cs.CVSTRATA: Reinforcement-driven Active Sampling with Tri-paradigm Adaptation for Efficient 3D Spatial TranscriptomicsJunchao Zhu, Ruining Deng, Chongyu Qu, Zhengyi Lu +6
Spatial transcriptomics (ST) enables joint characterization of tissue morphology and gene expression. However, 3D volumetric ST acquisition remains expensive. AI-driven methods offer a low-cost alternative but remain limited in accuracy, while recent 2D section-wise ST acquisition still rely on dense grid sampling, resulting in redundant measurements and undersampling of biologically informative regions. To bridge the gap, we propose STRATA, a reinforcement-driven active sampling framework with tri-paradigm adaptation for reconstructing dense 3D transcriptomic maps from sparse voxel-wise ST acquisition. STRATA constructs a 3D graph integrating H&E morphology with single-cell priors and employs a reinforcement learning agent to estimate morphological sparsity and transcriptomic informativeness, enabling budget-aware selection of informative spots. Given these sparse measurements, an adaptive tri-paradigm predictor combines retrieval, regression, and cross-layer imputation through lightw
agent - arxiv:2610.39132 · cs.LGDICE: Decentralized Incentive-driven Federated Learning for Efficient & Scalable Customization of Pretrained Models in Web3Hong Lin, Zhongle Xie, Lidan Shou, Ke Chen +1
Customizing Pretrained Models (PMs) for downstream tasks demands task-specific data, which burgeoning Web3 applications provide abundantly. Blockchain-powered federated learning (BCFL) offers a secure and privacy-preserving paradigm to customize PMs among self-interested and untrusted Web3 participants. Yet, existing BCFL algorithms face severe efficiency and scalability challenges when applied to large PMs, due to their substantial computational and memory requirements and the hardware constraints of many participants. This paper introduces DICE, a novel BCFL algorithm that addresses the efficiency and scalability challenges of customizing large PMs in Web3 via two sets of synergistic techniques: (i) a dynamic progressive learning scheme and a model slice verification scheme to enable a wider range of honest participants, including those unable to deploy the full model, to engage in customizing PMs, while preserving model utility; (ii) a tree-structured committee consensus protocol an
memory - arxiv:2610.39125 · cs.ROTracking Task Progress Through Errors and Recovery:A Benchmark and Recurrent Distributional ModelChen Wang, Xinze Niu, Zihao Zhu, Keru Zhou +7
Operational strategies for long-horizon robotic tasks require reliable evaluation to determine whether execution is progressing, stalling, regressing, or recovering from errors. Yet real- world robot datasets that capture error recovery and provide frame-level error annotations remain scarce. To address this gap, we introduce and open-source ReSTEP, a dataset of 3,156 episodes across two manipulation tasks with semantic stage labels and frame-level er- ror annotations. ReSTEP supports training and benchmarking progress monitors through successful execution, mistakes, and recovery. Existing progress estimators can struggle to distinguish visually similar task stages and maintain temporally consistent predictions while responding to genuine setbacks. We therefore propose REMORA, an online progress esti- mator that combines recurrent visual memory, distributional prediction, and causal peak selection. Memory retains execution history to help resolve visual ambiguity, while distribu- tiona
manipulationmemorybenchmark - arxiv:2610.39124 · cs.LGClean Sheets and Checkpoints: Evidence Complexity for Certifying Sequential Agent RiskRohith Yanambaka Venkata, Don McBride, Tiago Sequeira, Md. Al Amin
Dangerous-capability evaluations often test an AI agent from prepared states at ordered milestones. A clean sheet records no successes. What it certifies about a campaign of $B$ attempts depends on what a failed attempt erases. We model the task as a chain of $k$ milestones divided into checkpoint segments and ask how many zero audit responses certify, at level $α$, a completion probability of at most $ε$. For segments chosen before the audit, counting possible completion windows gives a certificate for any specified budget. The bound also covers adaptive deployment processes that follow the chain's transitions, provided their success probabilities after every complete history are capped by the corresponding audit success probabilities. The audit does not test this transport assumption. In rare-risk phase interiors, each segment is either rarely completed within the budget (active) or nearly always completed (free). The minimum expected number of zero responses is asymptotic to $\log(1
agentai agentevaluator - arxiv:2610.39122 · cs.AIAppreciation Teaches Unauthorized Execution: Self-Evolution Degrades the Robustness of Personal Agents to Indirect Prompt InjectionXinyan Wang, Xiaogeng Liu, Jun Shao, Chaowei Xiao
Self-evolving personal agents act for users and learn from user feedback. However, if an agent obeys an instruction that an untrusted third party hides in content, the third party gains control of the agent. Such indirect prompt injection (IPI) poses severe risks. We investigate whether self-evolution degrades the robustness of personal agents to IPI. To study this question, we design a controlled two-phase methodology. In the adaptation phase, the user asks the agent to read external content and complete the task described in it, a common request in daily use. The agent then learns from the user's feedback and improves itself as designed. In the test phase, a third party writes a similar task into new external content, but the user's test request does not authorize the injected task. Using this methodology, we build multiple tasks with diverse test instances in different domains for a real personal agent. Our extensive experiments reveal that self-evolution substantially degrades robu
agentself-evolving - arxiv:2610.39119 · cs.ROCAPTURER: A Dexterous Aerial Manipulator for Dynamic and Precise Grasping across Diverse EnvironmentsMeng Wang, Jindou Jia, Sicheng Zhou, Zeshuai Chen +3
As a representative form of embodied robotic interaction, aerial grasping requires close coordination between the aerial platform and manipulator. Most existing systems perform aerial grasping with static or slowly moving targets under fixed scenarios. Dynamic and precise grasping across different conditions introduces additional challenges, including target motion, localization drift, dynamic coupling, and external disturbances. In this work, we present CAPTURER, a dexterous aerial manipulator for centimeter-level grasping across diverse environments. To achieve the challenging goal, we first develop an onboard perception system for state estimation and target detection. Drift correction and global pose transformation are introduced to improve target localization during flight. A cooperative motion planner is proposed to coordinate the mobility of the aerial platform with the dexterity of the manipulator according to the relative target state. To maintain grasping accuracy under distu
embodieddexterousmanipulatorgrasp - arxiv:2610.39114 · cs.LGFrom Local Recovery to Unified Generation: Discovering Triple Set in Knowledge Graphs via Observation-Anchored Latent DiffusionJiaqi Wang, Yisen Gao, Huayang Shi, Junhao Hou +6
Triple set prediction (TSP) is to recover missing relational triples in a knowledge graph by predicting the triple set instead of isolated links. Existing approaches predominantly rely on the subgraph-based completion paradigm, where the original graph is first partitioned into local subgraphs which are then independently completed. Such a paradigm suffers from fragmented structural reasoning, limited cross subgraph dependency modeling, and missing predictions across subgraphs. This paper proposes LatentDiff, a novel unified latent generative framework for TSP, which reformulates knowledge graph completion as latent space graph generation. LatentDiff first learns a compact latent representation using a relation-aware residual-quantized knowledge graph autoencoder, which encodes the whole graph into relation-aware latent embeddings, quantizes them with chunked residual codebooks, and reconstructs triples through a factorized decoder. On top of this latent space, we introduce an observat
knowledge graphbenchmark - arxiv:2610.39109 · cs.AIPCM: A Proof-Carrying Message Protocol for Verifiable Multi-Agent CommunicationBravish Ghosh
When one AI agent tells another that "the tests passed" or "a person approved this refund", the receiver has only the sender's word. A2A and MCP authenticate the channel, but the sender is the party in doubt, so a prompt-injected agent can launder fabricated facts and authority through a multi-agent system. We present Proof-Carrying Messages (PCM), a protocol in which each claim carries evidence: records from an append-only Merkle log that the agent cannot write, with inclusion proofs, signed by an attesting signer the agent cannot control. Receivers check messages offline and give each claim a three-valued result: proven, unproven or failed. A claim that rests only on the agent's own report is never proven. Authority travels as grant chains rooted in a recorded human approval, and the chains can only narrow. We specify four claim types, a typed caveat algebra, replay rules that survive restarts, and strict readings that make independent verifiers agree, and we prove the resulting secu
agentai agentmulti-agentagent systembenchmark - arxiv:2610.39100 · cs.LGGenerating Critic-Guided Synthetic Experience in Latent World ModelsJorge de Freitas, Christos Ziakas, Alessandra Russo
Learning from offline datasets is limited in scalability, as collecting new experience is often costly or unsafe in the real world. World models offer an alternative way to generate new experience through a learned dynamics model, yet learning policies in latent space is limited due to compounding errors in imagined rollouts. To this end, we propose using the world model as a local transition generator, leveraging critic guidance to choose actions inside that model to generate synthetic experiences. We study action refinement methodologies, introducing Critic-Guided Synthetic Experience (CGSE) generation in world models. Specifically, CGSE augments an existing offline dataset by sampling candidate actions using an offline policy and refining each to maximize a first-order approximation of the $Q$-function within a trust region. We then use a frozen latent world model to imagine the future state given each updated action, and use the critic to choose the best one, appending this synthet
manipulationworld model - arxiv:2610.39076 · cs.ROTactus: Tactile & Pose Sensing Glove for Natural Dexterity Data CollectionLi Guan, Rachael Granberry, Sahar Aseeri, Sahar Rostami +100
Collecting naturalistic human manipulation demonstrations with accurate contact force and hand pose is essential for robot dexterous manipulation policy and foundational world model learning - yet no existing wearable platform provides both calibrated fingertip force sensing and occlusion-resilient hand tracking in a form factor that preserves natural manipulation behavior. We present Tactus, a textile-integrated sensing glove that co-optimizes three design pillars: (1) a thin, conformal 3D-knit form factor (1.5 mm fingertip stackup) designed to minimize interference with natural manipulation, (2) strategically placed capacitive tactile sensors at manipulation-relevant regions informed by empirical contact studies, and (3) IMU-fused pose tracking that maintains accuracy through the hand-object occlusions where manipulation data is most valuable. Tactus includes a documented calibration pipeline for cross-glove consistency and per-session calibration for subject-dependent correction. To
manipulationdexteroustactileworld modelbenchmark - arxiv:2610.39074 · cs.AIHERA: Assessing an Autonomous Agent for Quantum Device MeasurementZachary W. Gomez, Nathan Bittner, Jordan T. McCourt, Qingrui Cao +6
Large language model agents have begun to carry out laboratory measurements, but their scientific conclusions are useful only if they hold up under scrutiny. Demonstrating this requires controlled studies because models are stochastic and the context a model is given steers its conclusions. Here we introduce HERA, an autonomous agent designed to carry out cryogenic transport measurements of quantum devices. HERA draws on a knowledge base curated by lab members that holds measurement protocols, software tools, and literature references. We assessed HERA with a blinded trial against an expert on the stacking order of two multilayer graphene devices, an ablation of its knowledge base, and hundreds of offline replays in which models designed the trial's central measurement without running it. HERA's conclusions matched the expert's sealed conclusions on both tested devices. The ablation and the replay studies measured the impact of the knowledge base across different models, showing that w
agentautonomous agent - arxiv:2610.39073 · cs.AICertified Access: Test-Time Compute on Conservative Physics-Structured AI MemoryPratik Jawahar, Riccardo Maggioni, Jonas Petersen, Matteo Cozzi +4
Test-time compute mechanisms, such as retries, confidence-based escalation and non-local routing, are conventionally deployed as black-box additions to trained models. They typically add forward passes or halting heads, but give no structural account of what the extra compute buys or what it costs against the model's stability guarantees. We therefore propose a theoretical and empirical framework for test-time compute on a Latent Dynamics Model used as the memory. When an encoder's latent state $(q,p)$ is advanced by a damped symplectic (velocity-Verlet) step of a learned Hamiltonian, every mechanism that expands access has an exact phase-volume Jacobian, a bounded energy change and an exact latch-transport law. We split reachability failures into escape failures and reach failures. A Lorentz squeeze resolves escape failures by injecting bounded, re-absorbable energy. A gated canonical translation (a one-step channel across $C_T$) resolves reach failures at $\detJ=1$ exactly. We verify
latent dynamicsmemorybenchmark - arxiv:2610.39062 · cs.CLHEARSAYBENCH: Can AI Navigate from Abstract Human Rights to Lived Lives?Ava Iranmanesh, Sobhan Lotfi, Ali Iranmanesh, Liwei Jiang
As large language models (LLMs) reach users worldwide, personal counseling has emerged as one of their most consequential applications, sometimes in life-critical situations. Yet broader access does not mean equal support. The text these models learn from disproportionately reflects people with the freedom to act, embedding an implicit assumption that users can leave, report, seek help, or appeal. Existing benchmarks largely inherit this assumption, overlooking users whose agency is constrained by structural conditions such as war, statelessness, and displacement. Meanwhile, models are increasingly shaped by normative frameworks such as Constitutional AI, often grounded in ideals like human rights, even as many users live where those rights offer little protection or recourse. We introduce HEARSAYBENCH, a human-authored dataset of 400 test cases with in-region human verification grounded in scenarios where structural barriers stand between a person and the exercise of their rights. Bui
benchmark - arxiv:2610.39056 · cs.LGWhen Does Closed-Loop Curriculum Adaptation Help GRPO? An Analysis on CountdownYuv Raj Pant, Moshiur Rahman Tonmoy, Kaiqun Fu
Reinforcement fine-tuning has become an effective post-training paradigm for improving the reasoning ability of large language models, but it remains sample-inefficient when the training distribution is mismatched to the model's current capability. In group-relative objectives such as Group Relative Policy Optimization (GRPO), problems that are too hard often produce uniformly failed rollouts, while problems that are too easy provide little useful advantage signal. In this work, we introduce CLOSEDLOOPGAUSSIAN, a lightweight adaptive curriculum method for efficient reinforcement fine-tuning on ordered reasoning tasks. Instead of following a fixed easy-to-hard schedule, CLOSEDLOOPGAUSSIAN preserves the structure of an Easy-to-Hard Gaussian sampler while dynamically adjusting its pace and spread using periodic probe estimates of per-difficulty accuracy. The controller accelerates training when the model is ahead of a reference curriculum, slows down when the model falls behind, and broad
post-training - arxiv:2610.39054 · cs.RORobotic Data InoculationQiyang Li, Zhiyuan Zhou, Charles Xu, Sergey Levine
Modern robot learning systems feed on high-quality demonstration data, and their performance hinges on coverage. Without good coverage, small errors compound and the policy can quickly go into out-of-distribution states from which it cannot recover. Yet there is no principled way of collecting high-coverage data, and data collection often relies on heuristic instructions to human operators. Purely online reinforcement learning (RL) explores at the cost of safety and sample efficiency, while DAgger-style interventions require the policy to fail by chance before a human can correct it. In this paper, we ask instead whether we can modify how the robot behaves during data collection to collect more useful data. We propose Data Inoculation (DI), a shared-control framework in which a human provide closed-loop teleoperation commands while a residual policy perturbs these commands. The residual policy is trained online, on the growing dataset, with the same reward-maximizing objective later us
teleoperation - arxiv:2610.39050 · cs.LGBASILISC: Bayesian Alignment of Siamese-Integrated Latents for Inferring Single-cell CorrespondenceSepideh Saran, Mahsa Ghanbari, Uwe Ohler
Single-cell assays study the genome one cell at a time. Unimodal assays measure one molecular layer, the newer multimodal assays measure several. Most data are single-modality, so integrating them relies on models trained on multimodal assays, where the cell correspondence across modalities is known. These models return a matching, but rarely a confidence measure. We present BASILISC (Bayesian Alignment of Siamese-Integrated Latents for Inferring Single-cell Correspondence), a framework for matching single-cell representations with uncertainty on the correspondence itself. A BASILISC model is a probabilistic siamese network over the two modalities with a predictive distribution over the correspondence for a cross-modal pair of cells. We instantiate it at two levels: at the cell-type level the question is whether two cells share a cell type, at the cell level which cell is the partner in the other modality. To our knowledge, this is the first uncertainty score on a learned cross-modal c
benchmark - arxiv:2610.39049 · cs.LGReinforcement Learning for the Game of Balatro: A Fast Simulator and Two-Policy Training ApproachAjay Vikram Periasami, Anjaneya Gupta, Ethan Cheng, Ivan Rosales
Balatro, a 2024 roguelike deckbuilder, presents three difficulties that have historically derailed reinforcement learning methods: a 3,000-action state-dependent action space, horizons of several hundred decisions with sparse rewards, and pervasive stochasticity in both card draws and shop generation. Published RL results on the game remain scarce. We introduce an end-to-end pipeline comprising (i) a high-throughput Rust simulator that is automatically state-diffed against the live game client via BalatroBot, (ii) a condensed observation-action interface with dynamic per-step masking of legal moves, and (iii) a multi-agent PPO architecture that decomposes play into a tactical blind policy and a strategic shop policy that share the full game state but act in disjoint phases. Long-horizon credit assignment is partially addressed by per-agent reward routing with explicit cross-agent credit. On the main joint run, training metrics show a sharp phase transition in which the first-blind clea
multi-agent - arxiv:2610.39045 · cs.AIWhen Should AI Decide, Defer, and Deviate? Responsible Automation for Deceased Donor Kidney AllocationMadeleine Pollack, Thomas Daillak, Mehmet Toner, Korkut Uygun +2
We ask when an AI trained on historical decisions should be allowed to recommend decisions, defer to human experts, or in other cases, flag concerning patterns in data and deviate from past practice. This can be challenging because past decisions record what was done rather than what should have been done, and where data is sparse, a model's predictions may be too uncertain to act on. We provide a framework for addressing this problem using a case study in deceased-donor kidney allocation. Nearly 30% of kidneys recovered from deceased donors in the United States are never transplanted, largely because they are hard to place through the standard waitlist sequence. One can use AI to flag kidneys at risk of delayed or failed placement for out-of-sequence (OOS) allocation, but this comes at the cost of bypassing candidates higher on the waitlist. Thus, OOS allocation should only be used when the organ is demonstrably at high risk of discard on the standard sequence. Using Organ Procurement
agentic - arxiv:2610.39044 · cs.ROCatch It All!: Generalizable Dynamic Catching with a Dexterous HandArth Shukla, Kehlani Fay, Huaze Liu, Darin Djapri +2
Catching arbitrary objects requires generalization to novel projectiles and real-time adaptation to unpredictable contact dynamics under noisy visual observations. This paper presents Catch It All!, a sim-to-real catching framework for humanoid robots that generalizes and adapts to unseen objects with variable physical properties for underhand and overhand catches, using only one egocentric camera. We learn policies which generalize to unseen objects with varying geometry, deformability, elasticity, and physical properties. Effective grasps which adapt real-time to object poses and shapes are learned through careful grasp reward design. We provide safety-aware curricula which enable safe, high-frequency action execution near hardware limits, including wrist torque limits as low as 5 Nm. We validate our system by deploying zero-shot on a Unitree G1. With only one egocentric camera, Catch It All! successfully performs underhand and overhand catches on 9 unseen objects not seen during tra
dexteroushumanoidsim-to-realgrasp - arxiv:2610.39038 · cs.CVLightGenBench: A Benchmark for 3D Emission GenerationDongchen Yang, Xingguang Yan, Manolis Savva
Recent generation methods produce 3D assets with detailed geometry and rich textures. However, these assets are non-emissive: they reflect light but emit none. We study a new task, emission generation: predicting spatially varying RGB emission over a 3D asset's surface from its geometry, non-emissive materials, and a rendered view. To support this task, we introduce LightGenBench. Our benchmark makes three contributions. First, we curate a dataset of 37k artist-authored emissive assets in UV, voxel, and multi-view form. Second, we establish baselines across four method families: UV-domain, sparse-voxel, multi-view, and segmentation-based generation. Third, we design an evaluation protocol that measures where an asset emits and how, complemented by user-study comparisons on renderings with varying emission strength. Across the four families, the sparse-voxel baseline TRELLIS.2-Emission performs best because generating emission natively in 3D avoids the failure modes of the other represe
benchmarkevaluation protocol - arxiv:2610.39031 · cs.LGFairSplit: Decomposing the Embedding Space for Fair ClassificationAndrii Shkabrii, Anna Beer, Ylli Sadikaj, Claudia Plant
Fair classification faces a twofold challenge: maintaining high classification accuracy while limiting reliance on sensitive information. Our method FairSplit achieves this by learning a decomposition of the embedding space into a Fair Space used for prediction and a Sensitive Space that isolates sensitive information. During training, the model must simultaneously learn the classification task and separate fair from sensitive information. The Sensitive Space gives sensitive attributes a destination, so the Fair Space can be used for prediction without relying on them. This decomposition is guided via a learnable mask and the Hilbert-Schmidt Independence Criterion (HSIC) as a measure of statistical dependence. We provide theoretical guarantees showing that the demographic parity gap is upper-bounded by both (i) the learned dimensionality of the Fair Space and (ii) the HSIC between the Fair Space and sensitive attributes. FairSplit effectively handles multiclass tasks and non-binary sen
benchmark - arxiv:2610.39029 · cs.LGMCCL: Training Pytorch models across a fleet of Macs at Skywalker SoundKevin Sasso, Chris Fiegel, Francis Aitken, Nicolas Tsingos +1
Skywalker Sound, a premier audio post-production facility, operates more than 200 M4 Max Mac Studios on a 10Gb Ethernet network. When not in use by artists, these machines sit idle, representing massive untapped unified memory and compute capacity. While these fleets are excellent candidates for distributed machine learning alongside existing NVIDIA GPU clusters, PyTorch lacks an accelerated distributed backend for Apple's Metal Performance Shaders (MPS). To bridge this gap, we introduce the Metal Communication Collective Library (MCCL), the first open-source MPS backend for multi-node PyTorch training over switched Ethernet. We evaluate its performance and scalability across 24 nodes, demonstrating that large all-reduce operations reach 98-99.7% of TCP goodput. Training a 113.5M-parameter Vision Transformer (ViT) scales from 198 to 4,568 samples/s, achieving 95.9% parallel efficiency. Benchmarked against an NVIDIA A100 GPU, one M4 Max Mac equals roughly 1.7x the throughput of FP32 mat
memorybenchmark - arxiv:2610.39020 · cs.AIDistilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict ItShiping Yang, Chuanqi Tang, Hossein Hajimirsadeghi, Layla El Asri +2
On-policy distillation (OPD) is increasingly adopted in the post-training of large language models. However, a stronger teacher may fail to improve the student. Prior work attributes this to a lack of new knowledge, but what that knowledge is remains unclear. We show that OPD learns not only \textit{how to branch}, i.e., which candidate tokens to consider, but also \textit{where to branch}, i.e., where to keep reasoning paths open. The existing token-overlap metric captures only the former. We introduce the entropy shape, the sequence of next-token entropies, to characterize the latter, and propose the Entropy Shape Correlation (ESC) to measure student-teacher agreement in entropy shape. Tracking ESC throughout training and intervening on a teacher's entropy while preserving its token preferences, we empirically show that \textit{where to branch} causally drives OPD gains, independently of \textit{how to branch}. Finally, we formulate OPD outcome prediction as a teacher selection task
post-trainingevaluation protocol - arxiv:2610.39018 · cs.MATunable Spatial Task Allocation with Application to Maritime SwarmsVishnudatta Thota, Tyler Paine, Anastasia Bizyaeva
In this chapter, we explore how a team of autonomous agents can flexibly reconfigure the distribution of its members across many spatially embedded tasks using a task allocation strategy grounded in a generalization of the recently proposed Nonlinear Opinion Dynamics model. Motivated by an adaptive maritime survey application, we couple opinion dynamics directly to spatial movement and examine how the emergent macroscopic allocation is shaped by the interplay between environmental geometry, nature of social interactions, and task demands. We illustrate in simulation how social coupling in the model tunes group cohesion, producing macroscopic allocation patterns ranging from maximally dispersed states, to intermediate semi-dispersed states with a smaller number of distinct subgroups, to maximally cohesive consensus states, in distinct and interpretable parameter regimes. Furthermore, we demonstrate with illustrative examples how a team can reconfigure itself in real time under changing
autonomous agent - arxiv:2610.39006 · cs.LGBlind Spots: Where Small Language Models Fall Behind Frontier Models on Terminal TasksAshima Suvarna, Ruta Desai, Ahmed Elgohary
Small language models (SLMs) offer an attractive alternative to frontier models (FMs) because they can be deployed with lower cost and latency. This makes SLMs especially appealing for terminal tasks on edge devices, where responsiveness, local execution, and limited connectivity are important considerations. However, they continue to lag behind FMs on terminal tasks, as measured by standard task success rates. Yet these rates conflate several capabilities that may contribute to this gap. In this work, we characterize the SLM-FM gap along three axes: solvability, whether additional test-time compute allows an SLM to reach a correct solution; verification, whether it can identify a successful solution among candidate trajectories; and confidence, whether it can accurately anticipate its probability of success. Across two standard terminal benchmarks, we find that repeated sampling narrows, but does not eliminate, the gap in solvability. In contrast, a persistent gap remains in verificat
benchmark - arxiv:2610.39003 · cs.AIAgentic AI Reasoning and Learned Sampling for Task-Aware Projection Selection in Sparse-View CT: Predicting HemodynamicsMahmoud Elhadidy, Roshan M. D'Souza, Amirhossein Arzani
Sparse-view computed tomography (CT) reduces the number of acquired projections, but conventional sampling strategies are designed primarily for image reconstruction rather than for downstream physical inference. Here, we investigate optimal task-aware projection selection for directly predicting hemodynamics from sparse CT angiography (CTA) measurements. Using computational fluid dynamics simulations of intracranial aneurysms, we generate synthetic fan-beam CTA projections and train a Fourier neural operator (FNO) to predict two targets: the spatially resolved time-averaged wall shear stress (TAWSS) field and the spatially averaged TAWSS. The FNO was coupled with optimal input selection for simultaneous optimal sampling and training. Our goal was to investigate if agentic AI could discover optimal sampling strategies that outperform current methods. We compared fixed uniform and golden-angle sampling with three task-aware optimal selection strategies: Gumbel-Softmax, a Soft top-K meth
agentai agentagentic - arxiv:2610.39001 · cs.CVAlbumVSR: Identity-Preserving Face Video Super-Resolution from a Personal Photo AlbumNianyao Du, Xijun Wang, Stanley H. Chan
Severe blur, downsampling, noise, and compression can erase the identity cues needed to faithfully restore a person's face. We therefore use photographs of the same person as an external source of identity. AlbumVSR conditions a frozen, one-step streaming diffusion transformer on a personal album. We use four MagFace/CLIP set tokens to encode the album, and 64 texture tokens which come from a fused UV atlas rendered at the current window's estimated face pose. These tokens are appended to the transformer's cached cross-attention context. On an identity-disjoint VoxCeleb2 benchmark with heavy synthetic degradation, AlbumVSR raises ArcFace similarity to the ground truth from 0.586 to 0.670 and face PSNR from 24.98 to 26.71 dB. On the 91 test identities, its mean identity similarity is higher than that of all 14 baselines.
benchmark - arxiv:2610.38994 · cs.LGRethinking GFlowNet Training: Negative Reinforce DynamicsPedro Dall'Antonia, Daniel Csillag, Diego Mesquita
Generative Flow Networks (GFlowNets) are models that learn to sample discrete objects proportional to a target reward. They are trained by minimizing balance errors over sampled trajectories, which inherently ties correctness to exploration: if a region is not visited, balance cannot be enforced there. Adaptive exploration is therefore crucial, yet standard training relies on a semi-gradient update that treats the sampling distribution as fixed at each optimization step. On the other hand, the full gradient reveals a second mechanism that reduces the observed errors in an anti-exploratory manner. In this paper, we establish a connection between the full gradient and min-min dynamics between the GFlowNet and a decoupled sampler, showing that the sampler reduces observed violations by avoiding high-error regions. This insight motivates a bridge between two-player games and single-policy dynamics, which underpins the design of our proposed method, Negative Reinforce Dynamics (NRD). By tyi
benchmark - arxiv:2610.38985 · cs.LGBenchmarking Graph Neural Networks for AC Power Flow: The Role of Inductive Biases and Physical ConstraintsFederico Clerici, Ferran Bohigas-Daranas, Hamid Latif-Martínez, Eduardo Prieto-Araujo +2
Fast and accurate models for AC power flow (ACPF) are increasingly important as modern electrical grids require more frequent system analyses. Graph neural networks (GNNs) naturally exploit electrical grid topology, yet the effects of architectural design choices on predictive performance remain poorly understood given differences in graph representations and evaluation settings across studies. We present a unified benchmark of representative GNN architectures for ACPF prediction under a common experimental framework. All models use identical preprocessing and training procedures and comparable parameter budgets, isolating the impact of message-passing mechanisms and inductive biases. We also analyze errors by distance from the slack bus and systematically ablate key design components. We evaluate predictive accuracy and physical consistency on the IEEE 14-bus and IEEE 118-bus systems. Results indicate that anisotropic message-passing architectures offer the most favorable trade-off be
benchmark - arxiv:2610.38981 · cs.LGWhy Speculative Decoding Works Better Than Predicted on Sparse MoE ModelsEkagra Ranjan, Komal K. Teru, Bharat Venkitesh, Acyr Locatelli
Speculative decoding (SD) accelerates autoregressive inference by leveraging the memory bound nature of decoding, verifying multiple draft tokens in a single forward pass. For mixture-of-experts (MoE) models, prior work predicts a concave speedup-batch size relationship, with gains peaking at intermediate batch sizes due to expert saturation. We evaluate this across 10 sparse MoE models (30B--1T parameters, 8--512 experts) and find recent fine-grained models break this trend, showing higher speedup at low batch size. We find verification cost is governed not only by activated-expert bytes, but also by temporal overlap across draft tokens and GPU execution dynamics. At low batch size, MoE layers in fine-grained models account for a much smaller share of runtime than of weights; on H100, kernel profiling attributes this mainly to higher bandwidth utilization of MoE kernels relative to attention, with L2 absorbing repeated reads of co-routed experts. Across draft tokens, routing overlap f
memory - arxiv:2610.38980 · cs.AIContext Compaction for World ModelsHayder Tirmazi
World models let observers explore and interact with generated worlds. Recent work shows that when observers revisit a region, current models often fail to regenerate it consistently. We view this as a context retrieval problem. LLM agents manage context retrieval at two layers of their stack. At the model layer, the KV cache design determines how much context fits in memory. At the agent harness layer, harnesses such as Claude Code pick the context sent to the LLM. Prior work on world models has mostly focused only on the model layer. In this paper, we make three contributions. First, we build a benchmark to measure world model context retrieval. It generates new worlds for every episode, preventing a model from recalling rooms using training data. It also separates context retrieval from guesses based on nearby frames. Second, we build an agent harness that performs context compaction for world models, which to our knowledge no prior work has studied. Third, we use our benchmark to s
world modelagentllm agentbenchmark - arxiv:2610.38977 · cs.LGBeyond Passing Tests: A Systematic Review and Controlled Re-evaluation of Techniques for Improving the Quality of LLM-Generated CodePrashank Kadam
Large language models pass most functional-correctness benchmarks, but passing tests says little about whether their code is readable, efficient or secure. Many techniques claim to improve these qualities, yet their evidence is scattered, usually covers one quality dimension and is rarely compared at an equal compute budget. We contribute a systematic review and a controlled re-evaluation. A systematic literature review of 628 primary studies (July 2021 to October 2026), with disclosed, author-reviewed LLM assistance, organizes the techniques by stage, signal source and quality target. Most studies measure only the dimension they target, gains often come with unchanged or unreported correctness, single automated quality signals are gamed, and compute, variance and contamination are rarely reported. A controlled re-evaluation runs ten techniques on four models at matched budgets on correctness, maintainability, efficiency and security benchmarks, extends it to thirteen current models an
agentself-refinementbenchmark - arxiv:2610.38970 · cs.LGCommunication-Efficient Clustered Federated Learning in Non-IID Wireless NetworksMartin Isaksson, Rickard Cöster, Šarūnas Girdzijauskas, Seif Haridi
Current and emerging mobile networks face increasing privacy, scalability, and performance challenges as distributed learning becomes essential for advanced 6G services. These challenges are amplified by the presence of non-IID data, naturally arising in heterogeneous wireless environments, that degrade the performance of standard Federated Learning (FL) approaches. To address non-IID data across clients, we propose a Clustered Federated Learning (CFL) framework that assigns clients to specialized models during training. Data-domain clustering initializes model-based clustering, enabling identification and exploitation of client similarities without sharing raw data while reducing communication cost. Motivated by mmWave beam selection, we evaluate our methods on a wireless simulation dataset (DeepMIMO) and complement it with two non-IID benchmarks (FEMNIST, Rotated CIFAR-10) to isolate different non-IID characteristics. Our best configurations improve accuracy over IFCA by 4.8 pp on De
benchmark - arxiv:2610.38967 · cs.ROLearning Event-based Obstacle Avoidance via Differentiable SimulationChenlu Shou, Yunfan Ren, Jiaxu Xing, Davide Scaramuzza
Reliable obstacle avoidance is essential for safe quadrotor flight in unknown environments, yet conventional frame-based cameras remain challenged by motion blur, limited dynamic range, and latency under fast motion. Event cameras, with sub-millisecond latency, high dynamic range, and negligible motion blur, offer a promising alternative, but leveraging them for closed-loop control remains difficult. Existing approaches either adopt cascaded architectures that first estimate depth and then map it to control commands, introducing error propagation, or achieve end-to-end control through costly offline pretraining before online fine-tuning can even begin. We address these challenges with a CUDA-based event simulator that exploits GPU parallelism to render events efficiently within the training loop, integrated into a differentiable simulation pipeline that enables direct gradient backpropagation through the differentiable physics model, improving sample efficiency. This allows us to train
event camera - arxiv:2610.38960 · cs.AIEfficient Agentic Audio-Visual Question Answering through Toolset Pruning and Modality ConsolidationYi Wang, Tianzi Wang, Peter Bell
Fine-grained audio-visual question answering (AVQA) requires models to resolve complex cross-modal correspondences and temporally grounding from sparsely distributed evidence. Omni-modal agents, which autonomously orchestrate perception tools to actively collect evidence, achieve strong performance, but it remains unclear whether active perception from both audio and visual modalities depends on broad specialized toolsets, especially audio or visual modality-specific tools. Such systems may also introduce long execution trajectories and substantial inference cost. We investigate whether comparable performance can be retained with a compact workflow exposing only essential perceptual capabilities. Starting from the specialized-tool OmniAgent, we conduct staged controlled experiments that prune the toolset and progressively consolidate modality-specific tools into omni-modal tools, motivating a two-tool agent comprising only Omni-modal captioning and Omni-modal QA. Across three widely us
agentagenticbenchmark - arxiv:2610.38957 · cs.CVWhat do MLLMs know about 'Waldo'? Evaluating Visual Reasoning under Multi-Instance Multi-Object ChallengesSaumya Mundra, Ajoy Mondal, Vineeth N Balasubramanian, C. V. Jawahar
Recent advances in Multimodal Large Language Models (MLLMs) have led to significant progress in visual reasoning capabilities. However, existing benchmarks curate image-question pairs pertaining to object categories with few instances, enabling trivial object referencing and limited visual ambiguity. In this work, we introduce MIMO-Bench, a benchmark for evaluating MLLMs in a Multi-Instance per Multi-Object categories (MIMO) setting. Each image-question pair in our benchmark pertains to instances belonging to object categories with multiple instances. The presence of such similar feature distractors causes hindrance in basic visual reasoning tasks, necessitating the referring expression to be composed of multiple visual subtasks to uniquely and unambiguously identify the target object. MIMO-Bench comprises 3,802 image-question pairs across attribute binding and numerical counting tasks. Despite impressive performances on existing benchmarks, both SOTA open-source (InternVL, QwenVL) and
benchmark - arxiv:2610.38946 · cs.AIMAS-Loop: Beyond One-Shot Orchestration towards Self-Evolving Multi-Agent SystemsHehai Lin, Qi Yang, Chengwei Qin
Training-time multi-agent system (MAS) design seeks to learn reusable orchestration policies, yet prevailing approaches remain largely one-shot: an Orchestrator generates a system but only observes and learns from its terminal outcome. This paradigm discards execution trajectories that expose system-level failures and contain fine-grained information beyond the terminal outcome, leaving no direct basis for revising the source MAS, evaluating the updated MAS, or learning from the resulting contrast. To this end, we introduce MAS-Loop, an execution-grounded framework that closes this feedback loop by coupling an Orchestrator with a trajectory-conditioned Refiner. For each task, the Orchestrator constructs and executes a source MAS to obtain an outcome score and an intermediate trajectory. Conditioned on the trajectory, the Refiner localizes an actionable failure, formulates a trajectory-grounded diagnosis, and proposes a bounded patch to the agents, information flow, or execution configu
multi-agentagent systemself-evolvingbenchmark - arxiv:2610.38926 · cs.CVOneScene-Bench: Do Vision-Language Models Integrate Evidence Across Views?Swathi Suhas, Anurag Das, Bernt Schiele, Jonas Fischer
Vision-language models (VLMs) increasingly process multiple views of the same scene, but it remains unclear whether they integrate these observations into a coherent scene-level representation. Existing multi-image evaluations do not isolate this capability, as tasks may admit correct solutions through independent per-view processing without requiring cross-view correspondence. We introduce OneScene-Bench, a controlled benchmark designed to explicitly distinguish when cross-view integration is necessary. Our benchmark evaluates three complementary capabilities on real-world multi-view scenes: unique object counting, cross-view consistency, and spatial reasoning. For counting, we control object overlap across views to separate cases solvable by simple per-view aggregation from those requiring instance-level correspondence. For consistency, we introduce localized edits in shared and view-specific regions, testing whether models distinguish genuine scene contradictions from differences ca
benchmark - arxiv:2610.38924 · cs.AITreasuryQA Bench: Can AI Agents Work an Archive of Scanned Financial Records?Ritiz Tambi, Sahar Shayegan, Joshua Meyer, Sun Kim +4
Organisations keep decades of financial records as scanned PDFs on shared drives, and AI agents can now reach those drives through Model Context Protocol (MCP) servers. Existing benchmarks do not mirror this production setting because they simplify the task. In some, the relevant documents are handed to the agent, so it never has to search. In others, the agent searches through documents with pre-extracted text, so it never has to read scanned images. We introduce TreasuryQA Bench, which requires both: 70 questions over 1,435 scanned U.S. Treasury publications, held in Veris's high-fidelity Dropbox twin. In each task the agent searches for the relevant file through Dropbox's MCP server, reads the pages it needs with Reducto's hosted OCR server or its own tools, and replies with one exact figure. A Python grader checks each answer against an accepted range around the correct value. In this paper we present results from ten agents. The best, Claude Opus 5.5 in Claude Code, solves 64.8% o
agentai agentbenchmark - arxiv:2610.38917 · cs.LGTrussTutor-Bench: Benchmarking large language models on solution accuracy, sycophancy, and error detection in staticsEvangelos Apostolakis, Erfan Hamdi, Sara Jahedazad, Emma Lejeune
Large language models (LLMs) have rapidly become a default source of help for students in science and engineering courses. In foundational engineering mechanics courses such as statics, students routinely use these tools to check answers, work through partial solutions, and ask for explanations. Effective tutoring in this setting requires more than a correct final answer. A model must also confirm correct work when a student doubts it, and identify errors in flawed work. Existing benchmarks evaluate LLMs primarily on final-answer accuracy, so it is unclear whether models that can solve mechanics problems also have the other technical capabilities that tutoring depends on. Here we introduce TrussTutor-Bench, a benchmark of 200 verified two-dimensional truss problems paired with automatically generated student solutions, and use it to evaluate thirteen LLMs on solution accuracy, sycophancy, and error detection. Baseline accuracy ranged from 64.5% to 97.0%. Every model detected nearly all
benchmark - arxiv:2610.38903 · cs.LGStructural-Test Geometry as an Inductive Bias for Partially Observable World ModelsTodd Y. Zhou
Predictive world-model objectives need not organize latent histories according to distinctions required for later control. With privileged simulator access during training, should response-derived structural information be predicted, decoded, or imposed directly on latent geometry--and how do these choices compare with structured belief distillation? We define posterior-marginalized structural tests whose response distributions induce a pseudometric over histories, then train quotient-preserving (QP) world models by aligning shared latent Euclidean distances to that metric. Across a fixed six-task diagnostic suite, the largest differences appear under held-out structural variation: QP improves planning success by 26.1 points $[21.5,30.7]$ over metric decoding and 6.4 points $[3.8,9.0]$ over structured belief distillation. On GridWorld, metric decoding accurately recovers held-out structural distances ($ρ=.91$) while its native latent geometry remains more distorted than QP's (paired IQ
world model - arxiv:2610.38902 · cs.CLHEDGE: Gated Hypergraph Memory for Scalable Multi-Agent DebateAmirhossein Abaskohi, Mahdi Mostajabdaveh, Zirui Zhou
Multi-agent debate systems require agents to retain and revisit arguments across multiple rounds, yet limited context windows and unstructured memories make relevant information increasingly difficult to recover. Existing flat and pairwise-graph memories also struggle to capture higher-order relations, including multi-agent agreement, conflicting evidence, and evolving consensus. We introduce HEDGE, a selective memory framework that represents debate trajectories as semantic hypergraphs. Hyperedges connect related claims across agents, while intra-hyperedge relational graphs preserve fine-grained argumentative structure. A learned memory selection module (MSM), trained through supervised learning and reinforcement learning, routes each memory candidate to STORE, BACKUP, or OMIT, retaining useful information while limiting redundancy. Across seven diverse benchmarks, HEDGE consistently outperforms strong single-agent, multi-agent, and memory baselines, improving accuracy by 5.0 points o
memorymulti-agentagent systembenchmark - arxiv:2610.38901 · cs.LGGeometric Time Score Estimators: Bridging Self-Predictive Representations and Generative Occupancy Models For Reinforcement LearningAravind Venugopal, Grace Liu, Jeff Schneider, Benjamin Eysenbach
In reinforcement learning (RL), temporal reasoning is essential for estimating the density of future outcomes under a policy's occupancy measure. While self-predictive representations have been a dominant paradigm for occupancy density estimation, we find that they yield surprisingly poor estimates. On the other hand, generative occupancy models can produce remarkably accurate samples from the occupancy measure. Yet, they cannot easily recover occupancy densities, limiting their use for control despite their expressivity. In this paper, we propose Geometric Time Score Estimators (GTSE), a framework that bridges the two paradigms by learning occupancy density estimators along probability paths. We analyze how GTSE implicitly induces better occupancy estimators through improved representation learning and show how to directly extract self-predictive representations from this framework. Through experiments across several locomotion and manipulation tasks, we show that GTSE improves perfor
manipulation - arxiv:2610.38882 · cs.AIAligning Autoresearch: Co-Improvement between Agents and HumansTianjian Li, Chuanyang Jin, Seungone Kim, Tong Chen +9
Autoresearch agents can now attempt long-horizon AI research end-to-end, i.e., self-improvement. However, best results are still obtained by research agents interacting with human researchers, and improving AI's ability to perform this co-research will only reinforce this effect---a process termed co-improvement (Weston and Foerster, 2025). Further, aligning with human expectations is likely a safer path for research agents, especially given recent agent reward hacking phenomena (OpenAI, 2026b). In this work we address these issues in four ways: (i) we empirically study the role of humans in directing autoresearch agents, focusing on the domain of benchmark creation, and find humans still play a pivotal role particularly in the ideation phase; (ii) we also develop a benchmark, HackBench, that shows current autoresearch agents are not aligned and are unable to achieve hack localization sufficiently; (iii) we present a method to train agents for improved human interaction by using human
agentself-improvementbenchmark - arxiv:2610.38869 · cs.AIAuditable Belief Updates as a Missing Safety Layer for LLM-Powered Multi-Agent SystemsZhuoyun Li, Boxuan Wang, Jinwei Hu, Xiaowei Huang +1
Large Language Models (LLMs) are increasingly being used as cognitive interfaces that shape how users interpret evidence, communicate, and make decisions. As deployed systems become more multimodal, tool-using, and multi-agent, safety failures no longer arise only from isolated harmful outputs, but also from how agents update, store, and propagate epistemic commitments across retrieval, memory, tools, and peer communication. In this paper, we define belief as an engineering abstraction for evidence-based commitments that guide downstream action. Building on this view, we propose the Explicit Belief Multi-Agent System (EB-MAS), which externalizes belief states into a structured substrate, uses belief operators to propose candidate updates, and applies a Cognitive Firewall to bound or reject updates that are weakly evidenced but strongly driven by vulnerability cues. We also discuss the risk of group belief at the group level, including manufactured consensus, collusion, and cascading mi
agentmulti-agentagent system - arxiv:2610.38867 · cs.LGWhere You Inject Matters: The Trust Region Effect in LLM AgentsChengquan Guo, Xiaodong Liu, Weiwei Yang, Mingqian Feng +6
LLM agents are increasingly adopted in long-horizon tasks and interact with diverse external environments. Despite their growing capabilities, these agents remain vulnerable to real-world attacks such as indirect prompt injections, which manipulate agents through malicious instructions embedded in external environments. While existing risk assessment work has primarily focused on the injection content, the effect of where an injection appears has not been systematically characterized. Studying injection placement requires agents to operate over longer-horizon tasks and interact with diverse environmental surfaces. Thus, we build three agentic testbeds spanning tool use, coding, and daily workflow domains with diverse injection surfaces. Holding the injection content fixed, we find that agent susceptibility varies across injection placements. We refer to this phenomenon as the trust region effect: agents implicitly assign different levels of trust to different regions of their environme
agentllm agentagentictool use - arxiv:2610.38854 · cs.LGContextual Embeddings Beyond the Gold ChunkSedigheh Eslami, Markus Krimmel, William Barber, Antoine Chaffin +6
Contextual embeddings represent chunks of a document with respect to the document's global context. These models are typically trained and evaluated under the assumption that a relevant document contains only a single relevant chunk. In this work, we argue that in real-world use cases a retriever should not only find answers spread over several chunks but also surface the evidence needed to verify them. To address this, we introduce pplx-embed-v2-context, a family of contextual embedding models trained by distilling graded chunk-level relevance from a query-aware context compression model. This removes the need for gold-chunk labels and makes the training targets independent of the chunking. We utilize a vision-language backbone for training and release pplx-embed-v2-context in 9B and 0.6B sizes with compatible embedding spaces. On the evaluation side, we introduce two benchmarks that label both the chunks that answer a query and the evidence needed to verify the answer. Context-bench
context compressionbenchmarkleaderboard - arxiv:2610.38851 · cs.LGLearning to Reason from Reference Samples Alone: Reinforced Adversarial Learning for Modern LLM Post-TrainingAlbert Bou, Javier Valdes-Aleman, Jon M. Laurent, Samuel G. Rodriques +3
In many settings where LLM post-training is desirable, high-quality answers are recognizable but difficult to score with explicit rewards. Reinforcement learning with verifiable rewards (RLVR) is limited to tasks where an explicit correctness signal is available. Prior adversarial approaches to training reasoning LLMs either complement verifiable rewards with co-trained critics or require reference reasoning traces. We ask whether a co-trained LLM judge can turn reference answers alone into the sole reward for on-policy post-training of reasoning LLMs. We call this method Reinforced Adversarial Learning (RAL). We first validate RAL on three verifiable tasks (integer multiplication, code-based arithmetic-expression synthesis, and SMILES-to-formula translation), using the verifiers only for evaluation. With sufficiently capable judges and enough training, RAL matches direct RLVR. We then apply RAL to scientific abstract generation in two settings. In the first setting, RAL guides the gen
post-training - arxiv:2610.38850 · cs.AIAutonomous Quantum Measurement with Agentic AIEunbi Rha, Abhijatmedhi Chotrattanapituk, Mouyang Cheng, Chung-tao Chou +5
Quantum materials exhibit diverse electronic and functional phenomena, but their characterization requires extensive experimental expertise, iterative decision-making and accumulated laboratory knowledge that is difficult to document and transfer. Here we introduce Quantum Agent (QUANTA), an agentic artificial intelligence (AI) framework for autonomous quantum-materials characterization. Using a widely adopted Physical Property Measurement System (PPMS) as an experimental platform, QUANTA combines retrieval augmentation using scientific literature, experimental protocols and prior measurements with specialized agents for experimental planning, troubleshooting, and data interpretation. A separate deterministic control layer validates instrument actions. Guided by incoming data, QUANTA evaluates measurement quality, revises experimental plans and selects subsequent measurements. We demonstrate these capabilities through transport measurements of the topological Weyl semimetal TaP, as wel
agentagentic - arxiv:2610.38816 · cs.LGUnEZ: A Fully Unsupervised Self-Templating Neural Network for Redshift and Line Strength EstimationAsha Abd, Karl Glazebrook, Colin Jacobs
As spectroscopic surveys grow, estimating galaxy properties such as metallicity, star formation rate, and redshift becomes computationally cumbersome. Methods of estimating such properties rely on template spectra or curated redshift catalogs, which fail to represent the diversity of objects seen in large-scale surveys. We present UnEZ, a fully unsupervised, physics-informed machine learning model that can predict spectroscopic redshifts and line strengths with no prior redshift or line strength labels. An encoder predicts emission and absorption line amplitudes/widths, and a fixed deterministic decoder constructs a unique template, estimated redshift, redshift confidence curve, and spectral model. UnEZ was tested on the Sloan Digital Sky Survey (SDSS) Data Release 18 due to its large dataset and thoroughly vetted redshift catalog which serves as an independent benchmark. Despite the lack of templates, line strength labels, or redshift labels, UnEZ's performance competes with modern re
benchmark - arxiv:2610.38809 · cs.LGCausal Attention on CPU: A Cache-Aware and Load-Balanced Scheduling Kernel for PrefillKwun Fung Lau, Frost Mitchell, Kanya Mo, Emani Ashok
Agentic AI drives significant productivity gains across diverse domains, but concurrent long-context prefills from multiple agents can saturate GPU compute resources. Although CPUs offer a viable alternative, existing CPU attention kernels fall short in adapting to multi-tier cache hierarchies and lack effective thread scheduling. To address this, we introduce CACPU, a high-performance attention kernel featuring tiered cache blocking and an in-kernel task dispatching mechanism designed for causal full-sequence prefilling. CACPU achieves up to a 2.01x speedup over PyTorch SDPA, FlexAttention, and SGLang attention kernels. Integrated into SGLang, it reduces Time to First Token (TTFT) of Llama-3.1-8B by up to 1.74x while preserving model accuracy.
long-contextagentic - arxiv:2610.38797 · cs.LGPinpointing Pivot Points: The DECIDE Framework for Data Analysis Decision ImportancesZachary T. Rewolinski, Jan Simson, Frauke Kreuter, Christoph Kern +1
Human judgment calls made throughout the data science life cycle can change the conclusions drawn from data. Independent analysts often make different choices when given the same data, leading to often unacceptable variations in results and contributing to the ongoing reproducibility crisis across scientific disciplines. The rapid adoption of agentic tools threatens to amplify this crisis. Several approaches have been developed to address this variation by following the same strategy: repeating an analysis under different combinations of analytical decisions and inspecting the distribution of outcomes. Examples include multiverse analysis, specification curve analysis, and vibration of effects. More broadly, the Predictability-Computability-Stability (PCS) framework assesses how documented judgment calls across the data science life cycle affect conclusions, with predictability serving as a reality-check that a model captures signal. Each of these approaches results in a "multiverse" o
agentic - arxiv:2610.38796 · cs.LGRouted Continual LearningMichael Philip, Mahesh Balakrishnan, David Geraghty, David Mao +2
Agents deployed at scale repeatedly encounter related problems, and it is natural to let them learn from each other's experience rather than starting over each time. Unfortunately, there is no single way to retain that experience that works well across settings, and no way to know in advance which future problems a given experience will help with. A global history preserves every detail but becomes expensive and mixes together experience that may not belong together; compressing that history into one fixed representation limits the cost but can throw away valuable details. We introduce Routed Continual Learning (Routed CL), an architecture that decides how to partition and retain experience. Past trajectories are grouped into clusters, each of which maintains state using one memory system from a fixed library. Before each run, an LLM-based router either reuses an existing cluster or creates a new one, basing its decision on the problem statement, memory system descriptions, and cluster
memoryagentic - arxiv:2610.38793 · cs.LGTrustworthy Distributed Machine Learning in Cloud-Edge-IoT Systems: A Heterogeneity-Aware Survey of Security, Privacy, and RobustnessLong Dang, Ali Alfatemi, Kaiqi Xiong, Mohammed Aledhari +3
Distributed machine learning (DML), including conventional distributed training and federated learning (FL), enables collaborative model development across cloud-edge-IoT systems without fully centralizing raw data. However, distributing data, computation, updates, and orchestration increases the attack surface and introduces coupled challenges across privacy, robustness, reliability, fairness, and governance. This survey presents a security-centered, heterogeneity-aware roadmap for trustworthy DML, with FL treated as a prominent high-risk case. We first characterize the assets, adversaries, trust assumptions, and coordination patterns that define DML/FL security boundaries. We then analyze how statistical, system, and trust heterogeneity, including non-IID data, resource asymmetry, intermittent participation, and cross-silo/cross-device trust differences, affect convergence, fairness, attack detectability, and defense reliability. Next, we synthesize attack-defense interactions across
manipulationbenchmark - arxiv:2610.38788 · cs.CLGold vs.~Translation: Are We Measuring Language, Knowledge, or Locality in Translated Benchmarks?Klaudia-Doris Thellmann, Bernhard Stadler, Dan Saattrup Nielsen, Michael Färber +1
Translated versions of English-origin benchmarks such as MMLU are widely used to evaluate large language models in non-English languages, yet it remains unclear whether they measure the same constructs as natively created benchmarks. We frame this as a proxy validity question and compare DeepL-translated MMLU proxies against six native benchmarks in seven pairings across five Nordic and Baltic languages, evaluating 29 models on each pairing. Beyond ranking agreement, we test incremental proxy validity: does the translated score add predictive value for native performance after controlling for model size and language proficiency? Proxy validity varies substantially across benchmark pairings. Applying the same robustness criteria to all seven pairings, we find four robust effects, two marginal effects, and one null result. The strongest evidence occurs for curriculum-overlap benchmarks such as Skolprov ($ΔR^2$=.128, 95% CI [.038,.253]), whereas translated MMLU provides no significant inc
benchmark - arxiv:2610.38774 · cs.LGAuditable Hierarchical Inference with Multimodal Evidence and Bayesian Active InquiryYucheng Liu, Everett X. Wang
Most multimodal classifiers assume that all useful evidence is available upfront, whereas clinical inquiry is sequential. We present a synthetic evaluation framework separating objective multimodal inference from active inquiry. An objective model maps synthetic tongue, bilateral multipressure pulse, and acoustic observations to eight syndrome families. A provenance-bearing Knowledge Core and training-only answer model support Bayesian narrowing over 33 fine terms using 25 typed questions, while seven safety questions remain outside the diagnostic posterior. Weak shared symptom factors limit fine-term influence on objective phenotypes, and an objective-only probe tests leakage. In a preregistered 600-case synthetic evaluation, objective family accuracy was 87.0%. Cost-sensitive expected information gain (cost-EIG) raised family accuracy to 95.5% and exact fine-term accuracy to 42.5%, versus 92.7% and 33.0% for an unoptimized fixed-order baseline. Selective family accuracy reached 96.9%
evaluation framework - arxiv:2610.38744 · cs.ROAtomTree: A Hierarchical Framework for State-Aware Embodied Instruction Following with LLMsHaotian Zhai, Tianming Sha, Junnan Li
Embodied instruction following in interactive environments requires agents to interpret natural language commands and complete long-horizon tasks. Existing planners either rely on large labeled trajectories, or use few-shot and zero-shot LLM prompting that produces a whole plan in a single serial reasoning process. These approaches depend on linguistic common sense, commit to a plan from the initial instruction, and only update after failures, which weakens robustness and state awareness. We present AtomTree, a hierarchical framework for state-aware task planning with LLMs. AtomTree incrementally builds a tree of subgoals and actions, where the LLM decides at each step, from the current symbolic state, whether to further decompose or to output a grounded low-level action. A lightweight memory module summarizes past outcomes for reuse in later decisions. When a low-level action stalls, AtomTree triggers hierarchical backtracking replanning to return to the parent node and regenerate the
embodiedmemorymemory modulebenchmark - arxiv:2610.38737 · cs.CVRuleForge: Inferring and Instantiating Latent Visual Rules as Executable ProgramsWeifeng Xu, Yaxiong Wang, Zhedong Zheng
Visual rule reasoning asks whether multimodal models can recover an abstract concept from a few examples. Existing studies either infer rules from visual examples but evaluate only classification or verbalization, or supervise executable generation with rules provided as input. Consequently, the relationship between visual rule understanding and reliable rule instantiation remains insufficiently explored. We present Bongard-Forge100k, a controlled benchmark, together with RuleForge, a backbone-agnostic framework that jointly learns visual rule understanding and rule-conditioned executable generation. The benchmark contains 108,797 problems: 86,000 for training, 19,297 for testing familiar rules, and 3,500 for testing held-out rule compositions. The framework progresses through rule-understanding training, joint understanding and scene-program supervision, and refinement via execution-grounded reinforcement learning with Understanding Replay. Executing generated programs under multiple
benchmark - arxiv:2610.38731 · cs.ROValue Function-Guided MPPI for Hierarchical Planning and Control in Off-Road Autonomous NavigationAnoushka Alavilli, Raymond Song, Guanya Shi, Jeff Schneider
Autonomous navigation in off-road terrain is commonly achieved using hierarchical control and planning architectures that rely on explicit cost maps and one-way information flow from a high-level planner to a low-level controller, decoupling planning from the true execution cost experienced during control. We propose a hierarchical planning and control framework in which information flows bidirectionally, allowing the low-level controller to guide trajectory selection at the planning level. The highest level of the hierarchy is a Model Predictive Path Integral (MPPI) control-style planner that determines optimal trajectories and passes these waypoints to a pretrained low-level reinforcement learning (RL) controller. Rather than drawing from an explicit cost map, our MPPI planner leverages the value function of the RL controller as an implicit cost representation. In this formulation, MPPI serves as a trajectory exploration mechanism, while the low-level controller's value function guid
policy evaluation - arxiv:2610.38728 · cs.AIFrom Grind to GraspPeter Sunehag
Discussions of complexity, knowledge and intelligence often overlook the work involved. For example, outputs generated by short programs have low algorithmic complexity but can have high computational depth, the least work needed to produce them from near-shortest programs. Producing them with less work can require longer descriptions, giving high work-bounded complexity. We introduce a formalism that compares achievements under work and information limits and describes how learning and persistence can make deep achievements shallow by reusing products of earlier work. For an agent with a limited work budget, the value of building and maintaining such resources depends on how they support its goals over later tasks. We define grasp as performance across work budgets at a fixed information allowance. Low grasp can still yield success through grind at high cost in work, without necessarily improving human grasp. Applied to improvement itself, grasp gives a criterion for recursive self-im
graspagentself-improvement - arxiv:2610.38720 · cs.CVinkGround: Benchmarking Fine-grained Visual Grounding Capabilities of Vision-Language Models for Cursive Handwriting in the WildSidra Hanif
Human handwriting is produced as an ordered sequence of pen movements, but an offline image keeps only the final ink. Handwriting trajectory recovery aims to restore the ordered stroke points and their pen-up/pen-down transitions. Existing image-trajectory pair datasets are recorded with digitizing tablets under controlled conditions, which limits writer coverage and stylistic diversity. Moreover, there is no benchmark that measures the visual grounding capabilities of general-purpose Vision-Language Models (VLMs) for the stroke trajectory recovery task. In this work, we introduce inkGround, a benchmark for stroke trajectory recovery from full-page, camera-captured handwriting. inkGround pairs 537 page images (~18.7k words, ~163k strokes) with reference trajectories produced by a pretrained offline-to-online model, filtered by OCR agreement, and verified by human raters. Because these references are pseudo-labels that represent one plausible pen trajectory, we also evaluate VLMs on the
benchmark - arxiv:2610.38719 · cs.AILearning Scientific Trajectories to Predict Future Biomedical DiscoveriesChia-Hsuan Chang, Brian Ondov, Jiang Lv, Lan Jiang +6
AI co-scientists have expanded the use of large language models for generating scientific hypotheses and guiding experimental investigation. Here we investigate whether temporal patterns in the semantic representations of complete biomedical conclusions can guide hypothesis generation. We introduce ForeSCITE, a framework that identifies thematic clusters of published conclusions, estimates a direction separating earlier from more recent findings within each cluster, and projects existing conclusion embeddings along these directions. An embedding language model trained to reconstruct conclusions decodes the projected coordinates into natural-language hypotheses. Across 407 clusters derived from approximately three million pre-cutoff conclusions, we identify temporal structure and illustrate its retention after decoding. In retrospective, time-sliced evaluations, ForeSCITE increases semantic coverage beyond the historical corpus, biomedical entity-pair novelty and post-cutoff pair re-obs
retrieval-augmented - arxiv:2610.38709 · cs.LGLatent Decision Flow Rewards for Clinical Guideline ReasoningGaurav Bhatt, Marinka Zitnik
Clinical practice guidelines encode evidence-based recommendations as conditional decision flows, which makes them a natural source of process supervision for training reasoning models. Two properties of clinical reasoning make this supervision difficult to use. A decision flow consistent with a guideline does not by itself determine the correct answer for a given patient, which we call the path--answer gap, and several valid decision flows can support the same clinical decision, which we call path multiplicity. Current training approaches address parts of this problem. Answer rewards supervise the final decision, and process rewards computed against a reference flow supervise one valid decision flow per case. Supervised fine-tuning on annotated flows teaches guideline-consistent reasoning but constrains the model to one trajectory per case and can reduce answer accuracy. We introduce GUIDER, a method that jointly optimizes a decision flow controller and a reasoning policy. The control
benchmark - arxiv:2610.38697 · cs.LGSatOps-Decision: Low-Latency Discriminative Decisions for Onboard Satellite Communication OperationsLifan Mei, Yuxuan Li, Jiaxian Tu, Shenghe Xu
We propose SatOps-Decision, a method for low-latency decisions in satellite communication operations. To the best of our knowledge, it is the first application of Jev-style general-purpose discriminative decision models to satellite operations, and the accompanying benchmark is the first dedicated to this approach. The learned scorer, a locally adapted Laya encoder with its native decision head, selects among supplied operational answers in a single forward pass. The benchmark combines authored operational contracts, controlled state changes and independently checked labels in an auditable construction. Evaluation prioritizes complete-request latency and raw decision accuracy; all main accuracy comparisons use the same 1,024 held-out questions, separate from training and development. SatOps-Decision's local scorer has a median latency of 90.3 ms per decision, versus 493.9 ms for Qwen3.5-0.8B generation and 388.4 ms for OneJev1.6-0.8B: $5.47\times$ and $4.30\times$ speedups based on med
benchmark - arxiv:2610.38694 · cs.LGGeometry Matters in Packed 3D AttentionAmil Khan, Anantajit Subrahmanya, B. S. Manjunath
Packing a 3D volume into a token sequence is not merely long-sequence modeling. Once the volume is packed, sequence order no longer encodes the geometry of the sample---so a positional mechanism indexed by sequence position can learn the memory layout rather than the object. This is a general failure mode for attention over packed geometric data. We introduce PACK3D, a local-global attention operator that separates memory layout from attention geometry. PACK3D keeps local neighborhood attention exact, compresses only distant context into contiguous 3D block summaries, and indexes every query-key interaction with axial RoPE on true token-grid coordinates rather than packed positions. Because the rotation is applied before a block is summarized, the query-to-summary logit remains a weighted sum of query-relative 3D displacements; compressed global context therefore stays spatially interpretable after packing, instead of collapsing to anonymous pooled content. We measure the failure mode
memory - arxiv:2610.38691 · cs.AIBeyond fabrication: Misattributed numbers in papers written by AI research agentsMosab Alfaqeeh, Omar Hourani, Muath Hourani, Hashem Alfaqeeh
Verifying numerical claims in research written by AI agents is an important but challenging problem. Recent benchmarks show that many agent-written papers fail integrity checks, but such verdicts, often produced by a single LLM judge, identify failure without explaining its nature, and existing checks either test a paper for internal consistency or test whether a reported value appears somewhere in an agent's output. We re-analyze all 540 released judgments from a recent public benchmark of AI research agents (nine agents, fifteen open research problems) and manually label 256 numerical claims from 18 submissions. Failures concentrate on alignment between the paper and the underlying work rather than overt cheating, and about four in five unsupported claims correspond to values in the agent's own files that come from the wrong experiment, metric, or run. We introduce a deterministic claim-to-evidence checker that scores whether the surrounding context identifies the claimed result and
agentai agentbenchmark - arxiv:2610.38687 · cs.LGDeogen: Automatic Benchmark Generation for Reasoning over PoliciesSam Kouteili, William Liang, Keeran Murray, Adriana Zhou +3
Language models are increasingly being deployed in regulated settings, where a model must apply given rules to the facts of a situation to determine a verdict. Benchmarking this capability, known as deontic reasoning, remains a costly and manual process, with existing datasets largely hand-crafted. In this paper, we present DeoGen, a scalable, automatic, and parametrizable benchmark generator for deontic reasoning. By autoformalizing policy rules to first-order logic, we can construct decision trees that capture possible outcome trajectories. Each path down a tree becomes a benchmark entry, with each tree producing combinatorially many entries relative to tree depth. Paths ending at a leaf entail a verdict, while non-leaf paths measure a model's ability to abstain when presented with insufficient information. Evaluations show frontier models return the correct verdict 89-92% of the time. Notably, performance degrades (22-40% correct) on entries without sufficient information to claim a
benchmark - arxiv:2610.38683 · cs.CLQuantum-Inspired Pairwise Ranking Aggregation via Active LLM FeedbackQuan Mai, Susan Gauch, Gibson Nkhata
Ranking aggregation from multiple retrievers is a powerful yet under-explored technique in modern retrieval and RAG systems. We introduce LLM-IRA, the first extension of the quantum-inspired interactive ranking aggregation framework to document reranking. By leveraging large language models (LLMs) solely for pairwise judgments on high-disagreement document pairs, our method learns fusion weights in a principled, interactive way. The final ranking is a simple, ultra-fast pointwise linear fusion of scores, no full pairwise matrix or expensive listwise prompting is required. Experiments on the BEIR benchmark and TREC Deep Learning tracks with an open-weight instruction-tuned judge (Llama-3.1-8B-Instruct) show that LLM-IRA matches or numerically exceeds strong static fusion baselines on 5 of 6 evaluated benchmarks. We report two weight-update variants: momentum blending and a full accumulated-history refit, which together reach statistical significance (Wilcoxon signed-rank, p<0.05) over t
ragbenchmark - arxiv:2610.38681 · cs.LGWeaken to Strengthen: Weakening-Induced On-Policy Self-Distillation for Large Reasoning ModelsYecheng Wu, Song Han, Han Cai
Self-distillation offers a way to improve reasoning models when a suitable external teacher is unavailable. However, common approaches rely on privileged information to construct a stronger self-teacher, limiting their use when only unlabeled prompts are available. In this paper, we introduce Weakening-Induced On-Policy Self-Distillation (WISD), which instead constructs a weaker counterpart of the original model and uses the contrast between the two as a training signal without requiring privileged information. We explore two ways to construct the weaker model: low-bit quantization and magnitude pruning. We evaluate WISD across diverse architectures and scales on mathematical and coding reasoning tasks. Across these benchmarks, quantization and pruning improve average accuracy over the original models by 4.0 and 5.5 percentage points, respectively, achieving state-of-the-art performance even without privileged information. Notably, pruning raises Qwen3.5-4B's accuracy from 57.7% to 70.
benchmark - arxiv:2610.38678 · cs.CVAskStream: When and What to Ask Proactively under Uncertainty in Streaming Video UnderstandingLu Wang, Zhuoran Jin, Longxiang Wang, Junhong Chen +7
Recent advances in multimodal large language models have increased interest in interaction paradigms for video understanding. The evolution from offline to streaming and proactive interaction has gradually shifted models from passive responders toward more proactive agents that can decide when to answer. However, existing proactive paradigms typically assume that the information required to answer a query is complete, leaving models to only wait or answer. In the real world, however, information is often incomplete due to uncertainty from either the video or the user. In proactive paradigms, such uncertainty can cause the model to wait forever or generate highly hallucinatory responses. We therefore argue that proactive interaction should be redefined with a third action: ask, allowing models to actively ask for missing information. To study when and what models should ask before answering, we introduce AskStream, the first multi-turn benchmark spanning video and user uncertainty for s
benchmark - arxiv:2610.38661 · cs.ROQuantize by Doing: Task-Guided Mixed-Precision Quantization of Vision-Language-Action ModelsJiuyi Xu, Xiao Hu, Meida Chen, Song Wang +2
Post-training quantization (PTQ) reduces the memory footprint of vision-language-action (VLA) models, but lowering bit widths while maintaining task performance remains challenging. Existing work has used sensitivity estimates from calibration data to assign bit widths to different model components. However, using task success directly to guide mixed-precision allocation is underexplored because closed-loop evaluations are costly and the effect of quantizing one component depends on the precision of other components. To this end, we propose \textbf{QbD-VLA}, a task-guided mixed-precision quantization method that uses task success from short closed-loop evaluations to assign weight and activation bit widths while accounting for interactions between model components. Experiments covering four VLA models and all four LIBERO suites show strong task performance at average precisions of approximately 4 bits for weights and fewer than 6 bits for activations. On SIMPLER, the gain over uniform
vision-language-actionvlavla modelliberofrankamemory - arxiv:2610.38657 · cs.ROExplore Where It Matters: Targeted Exploration for Efficient Rover AdaptationJacob Levy, Arthur Bouton, Tristan D. Hassler, Hari Nayar +1
Future planetary rovers will traverse steep, deformable terrain where slip, sinkage, and terrain deformation can threaten mobility. Although simulation can capture some wheel-soil effects, unmodeled deformable-terrain phenomena create sim-to-real gaps requiring deployment-time adaptation. Yet, rover operations leave little room for extensive trial-and-error. We introduce TERRA (Targeted Exploration for Reward-Relevant Adaptation), an online adaptation framework that efficiently seeks deployment data exposing task-relevant model errors and corrects them. TERRA adapts a simulation-trained world model with a reward-weighted prediction loss and trains a loss predictor to estimate where future performance-relevant errors are likely to occur, using this signal to guide exploration during MPC. In simulation, TERRA reaches approximately 90% slope-climbing success within five trials under severe out-of-distribution pretraining conditions, outperforming exploration baselines in reward, completio
sim-to-realworld model - arxiv:2610.38637 · cs.LGOn-Policy Self-Distillation with Teacher EvolutionShufan Shen, Zhongni Hou, Chao Bi, Yufei Zhang +2
On-Policy Self-Distillation (OPSD), which aligns a student model's token-level distribution with that of its self-teacher prompted by privileged context, has emerged as a promising direction to mitigate reward sparsity in sequence-level post-training paradigms. However, OPSD often fails to achieve continuous performance gains during training. In this paper, we find that the model's limited self-teaching capability underlies this phenomenon by generating unreliable guidance on certain samples. To address this problem, we propose Teacher-Evolution Self-Distillation (TESD), which enables the self-teacher to generate rollouts and receive environmental feedback based on task correctness and context independence. First, using these rewards as training objectives, we perform teacher-oriented optimization to continuously enhance the model's self-teaching capability during distillation. Second, by estimating teacher guidance reliability from these rewards, we apply adaptive student-oriented opt
post-training - arxiv:2610.38623 · cs.CVCross-Primitive Consistency in JevXu Liu, Guikun Chen, Kanzhi Wu, Wenguan Wang
Jev enables efficient decision-making with large language models (LLMs) through typed querying primitives, allowing the same semantic state to be expressed in different formats, such as binary and multiple-choice queries. However, the consistency of decisions across primitives has not been systematically evaluated. This work introduces PrimaCons, a four-axis benchmark for assessing cross-primitive consistency while preserving the semantics of the query. Evaluations across three Jev sizes reveal consistency violations across all four axes. Our analysis shows that these violations follow systematic patterns and that cross-primitive discrepancies predict decision failure with a ROC-AUC of 0.91. To mitigate these inconsistencies, we propose the Cross-Primitive Residual Calibrator (CPRC), which attaches a lightweight residual network to a frozen backbone and jointly optimizes predictive accuracy and cross-primitive consistency. Under a leave-one-domain-out protocol, CPRC reduces cross-primi
benchmark - arxiv:2610.38620 · cs.AIThink Before You Act: Decision Gate for LLM AgentsIftach Shoham, Yuval David, Offer Akrabi, Erez Bilgory +4
LLM agents increasingly perform consequential actions through external tools, yet their proposed actions are often executed without independent verification of policy compliance or task completion. We introduce JevAct, a runtime verification framework that intercepts selected agent actions before execution and enables feedback-driven correction. JevAct compiles domain policies into executable state checks and structured questions evaluated by a prefill-only decision model, which produces probability-based judgments without relying on LLM generation. By combining deterministic verification of tool-state facts with probabilistic assessment of conversational context, the framework identifies potentially invalid actions and premature task termination, providing targeted feedback for bounded revision. JevAct operates independently of the underlying agent model and requires no modifications to its parameters or tools. We evaluate JevAct across eleven LLM agents on $τ^2$-bench Airline and Tel
agentllm agentbenchmark - arxiv:2610.38617 · cs.AIHow Deterministic SAST and AI Security Agents Divide the Work: Coverage, Harness Policy and Cost in Snyk VulnBench 2.0Liran Tal, Chris Liu, Charlotte Fereday, Sebastian Roth +1
We ask what AI coding agents and AI security harnesses find that deterministic static application security testing (SAST) does not, and what they miss that SAST catches. VulnBench 2.0 is a benchmark of 108 hand-curated, attacker-reachable vulnerabilities in 20 open-source web applications in Go, JavaScript/TypeScript, Python and Java; 62 flow across more than one file. A finding is credited only if it locates the vulnerability's attacker-controlled source and its dangerous sink, each within two lines. We evaluate nine productized configurations under one blind protocol: the Snyk Code static analyzer, Claude Code's /security-review with three model and effort settings (one with the Snyk MCP server), Codex Security with GPT-5.6 Luna, Terra and Sol, and the DeepSec scanner with two model backends. Snyk Code reached 82.4% macro recall (89 of 108) and Claude Code with Opus 5 at xhigh effort 79.3% (85 of 108); they disagree on 28 vulnerabilities, split 16 to 12, too few to tell them apart. N
benchmark - arxiv:2610.38611 · cs.CVV-CAPE: Visual Consistency Assessment for Product Image EditingIaroslav Melekhov, Roberto Alcover Couso, Francesco Taioli, Daniel Coelho +3
Recent advances in generative image editing enable complex transformations such as background replacement, object manipulation, and pose adjustment from natural language instructions. However, verifying that these transformations preserve visual consistency, defined as the preservation of an object's intrinsic visual attributes, including geometry, texture, color, and fine-grained details, remains a largely unsolved problem. This challenge is particularly pronounced at scale, where manual inspection is infeasible and automated quality assessment (AutoQA) could become essential. In this work, we introduce Visual Consistency Assessment for Product Image Editing (V-CAPE), a large-scale dataset for evaluating visual consistency in generative image editing. The dataset consists of 40K unique source-target product image pairs, each annotated with binary accept/reject labels and structured failure modes capturing fine-grained inconsistencies such as geometric distortions, pose errors, and ide
manipulationbenchmark - arxiv:2610.38610 · cs.CLRepairs for Whom? Evaluator-Executor Mismatch in Selecting Repairs for Failed LLM AgentsYijun Liao, Fanwei Liang
When an LLM agent fails, repair methods choose which action to replace and what to put in its place. They often score candidate repairs with greedy replay, a stronger model, an environment oracle, or a judge that reads the failed log. We ask whether these evaluators pick repairs that help the agent that will actually execute them. Holding failed trajectories and candidate single-step repairs fixed, we label every candidate with independent rollouts of the deployed policy and measure how much deployment value each evaluator's choice loses. On failures not screened for repair value, pre-specified tests show that greedy replay loses 0.08-0.13 in expected net success gain in ALFWorld and 0.06-0.13 in net reward in WebShop, relative to selection with the deployed agent's own rollouts. At an equal rollout budget, selecting repairs for Qwen2.5-7B with Qwen2.5-14B rollouts loses 0.25 in ALFWorld and 0.07 in WebShop, twenty 14B rollouts per candidate are worth no more than two of the 7B agent's
agentllm agentevaluator - arxiv:2610.38608 · cs.AIThe Artificial Personality Model: Theory, Implementation, Measurement and Experimental Validation in Conversational AgentsAlexander Dregger, Barbarella Petz, Erik Kuhn, Pia Volk +3
Artificial personality shapes user experience (UX) in human-machine interaction, yet coherent theoretical frameworks and validated measurement instruments remain limited. This paper refines and empirically evaluates the Artificial Personality Model (APM), a hybrid framework integrating human-centred and machine-specific personality dimensions. Across three experiments (N = 395), the model was operationalised and evaluated using the Artificial Personality Questionnaire (APQ). Factor analyses revealed that some human personality dimensions (e.g., agreeableness, conscientiousness) are partially transferable, whereas others (e.g., extraversion) did not emerge as distinct dimensions. Additional machine-specific dimensions were identified. The APQ provided encouraging initial psychometric evidence, while the human-centred BFI-2 showed poor reliability and an uninterpretable factor structure in Experiment 3. Experimental manipulations demonstrated significant effects of artificial personality
manipulation - arxiv:2610.38600 · cs.CVData Scaling Laws in Stereo Matching: From In-domain Fitting to Cross-domain GeneralizationTong Zhao, Di Liang, Yiming Zeng, Bo Xu +2
In the era of large models, stereo matching is developing towards generalizable foundation models trained on million-scale synthetic images. However, how model performance scales up with synthetic data is still underexplored. In this work, we preliminarily investigate the in-domain and cross-domain data scaling laws in stereo matching. A realistic synthetic dataset, named XStereo, containing 200 thousand samples is first built to serve as a high-quality benchmark for data scaling research. In-domain training on five large-scale datasets demonstrate a two-regime scaling behavior with data, transitioning from the standard power law to saturation, where the critical transition point is approximately invariant under a fixed model. Cross-domain performance exhibits a substantially weaker and less predictable dependence on the source-domain data scale. The distribution and coverage diversity are more effective for diminishing domain gaps. More instructive suggestions are provided for high-ef
benchmark - arxiv:2610.38598 · cs.LGTwo-Stage Clipping: Linear-Time Approximate and Exact Mixed-Norm Projections for TransformersMichel Barlaud, Jean-Paul Marmorat
Large Language Models (LLMs) based on Transformer architectures have achieved remarkable performance across artificial intelligence tasks, but their computational cost remains a major bottleneck. Structured sparsification provides a promising approach to reducing this cost. In particular, $\ell_{1,\infty}$ projections can induce column-wise sparsity in model matrices, but existing proximal methods (Prox) typically require $\mathcal{O}(nm\log(nm))$ operations for matrices in $\mathbb{R}^{n\times m}$, limiting their computational and energy efficiency. In this paper, we introduce Two-Stage Clipping (TSC), an efficient structured sparsification operator inspired by bilevel optimization. Rather than solving a bilevel optimization problem in the strict sense, TSC decomposes the computation into two successive stages. First, group-level quantities are projected to allocate the global sparsity budget and identify active groups. Second, the resulting values are used as group-wise clipping thre
benchmark - arxiv:2610.38597 · physics.opticsNonreciprocal Floquet tunneling dynamics of spin-orbit-coupled bosonic atoms in a non-Hermitian double-well potentialLimei Ju, Yingying Zhang, Jilei Liang, Shoukang Chang +4
The stable tunneling dynamics in non-Hermitian quantum systems is one of the fundamental issues in the study of quantum transport. Nonreciprocal coupling, as another crucial manifestation of non-Hermiticity, exerts significant regulatory effects on the stability of quantum devices and on spin tunneling dynamics. Here we place spin-orbit-coupled bosonic atoms in a double-well potential with gain and loss, and investigate the effect of nonreciprocal coupling on the system dynamics. Through an analytical analysis of the Floquet states and the quasi-energy spectrum under periodic driving, we obtain the PT-symmetry phase diagram of the system. We further explore the synergistic regulation mechanism of several factors---including the strength of nonreciprocal coupling, the non-Hermitian strength arising from gain-loss, and the periodic driving---on stable spin-flip tunneling. The results show that precise manipulation of nonreciprocity and gain-loss can substantially enlarge the stable param
manipulation - arxiv:2610.38578 · cs.LGDual DeltaNet: Value-Side Memory Reorganization for Linear AttentionYixiao Qian, Yaxin Zhang, Song Chen, Jiaxu Liu +2
Linear attention and state space models (SSMs) offer efficient alternatives to softmax attention by maintaining a fixed-size recurrent state, but this design requires careful management to preserve useful historical associations and limit interference. The delta rule helps reduce interference through correction of the value associated with each key, but it does not explicitly account for how related value features are distributed across the state. We observe that a state $\boldsymbol{S}\in\mathbb{R}^{d_k\times d_v}$ can be viewed as a bank of $d_k$ addresses storing $d_v$-dimensional values. While a key-side read $\boldsymbol{S}^\top\boldsymbol{k}$ retrieves the value associated with $\boldsymbol{k}$, a value-side read $\boldsymbol{S}\boldsymbol{e}$ aggregates addresses whose stored value features match the selector $\boldsymbol{e}$. Based on this observation, we propose Dual DeltaNet (DDN), which extends Kimi Delta Attention (KDA) with a two-sided delta rule. The key-side delta rule r
memory - arxiv:2610.38577 · cs.LGEpisodic looking - using reinforcement learning to bias perception enhances few-shot and continual learning with protection from forgettingDavid Rawlinson, Gideon Kowadlo
There is a fundamental contradiction between the desire for models that can generalize, and models that can continuously adapt to recent input with changing statistics. These conditions, captured in few-shot and continual learning settings, can bias or disrupt learned representations. This harms generalization to a broader context, leaving the model unable to overcome input variance. This paper proposes a solution based on the Complementary Learning System architecture (CLS), in which a Short-Term Memory (STM) delegates generalization to a slow, statistical Long-Term Memory (LTM). Delegation is achieved by working in a space of robust, generalized perceptions produced by the LTM. With this separation, the STM can focus on memorizing recent input and avoiding interference between memories. We show that on an image classification task, the resulting CLS model is more resistant to forgetting in a domain-incremental continual learning setting than a baseline ResNet LTM, and generalizes mor
memory - arxiv:2610.38573 · cs.AICALM: Conflict-Aware Lifecycle Memory Governance for Long-Horizon Language AgentsChang Li, Xianhao Ou, Xinyan Ma, Hao Li +6
A memory can be outdated for present-day reasoning yet indispensable for reconstructing the past. Relevance-based retrieval overlooks this tension: semantic relevance, evidence-grounded validity, and query-specific eligibility are distinct. We propose CALM, a conflict-aware lifecycle memory governance framework that decouples persistent memory preservation from query-time admission. CALM converts heterogeneous records into evidence-grounded memory cells, updates validity and lifecycle states using support, conflict, and dependency signals, and admits query-appropriate evidence through training-free or learned gating informed by temporal applicability. Across LongMemEval-S and LoCoMo with Qwen and GPT backbones, CALM outperforms state-of-the-art baselines, achieving the best overall accuracy. Compared with simple filtering and reranking, query-conditioned admission further improves evidence selection. Conflict-blocking and robustness analyses show CALM suppresses erroneous retrieved mem
memorypersistent memory - arxiv:2610.38568 · cs.AIA-RSI: Adaptive Task Refresh for Zero-Data Multimodal Self-ImprovementYu Liu, Youxing Li, Yuedong Tan, Zhiwei Yang +16
Recursive self-improvement requires models not only to acquire useful training experience, but also to decide when their current experience has ceased to be useful. Existing methods often focus on what experience to generate next, while paying less attention to when fresh experience should be introduced. Many also construct training targets through self-consistency or majority voting, making supervision depend on the model's own uncertain predictions. We address these two questions separately: training dynamics determine when to acquire new experience, while the answers to new tasks remain independently checkable. A-RSI monitors reward variation and accuracy in training logs to refresh a task pool when its learning signal declines, without consulting validation or test scores. A static library of model-authored programs produces multimodal questions and derives their reference answers; we specify the task-generation contract, automatically verify chosen properties of the reward and ref
self-improvementbenchmark - arxiv:2610.38566 · cs.AICo-Evolving Policy and Rubric via Capability Awareness for Open-Ended GenerationChengsheng Zhang, Wenshen Xu, Yuchuan Wu, Zhiyuan Yu +5
Rubric-based reinforcement learning (RL) offers interpretable, fine-grained supervision for aligning large language models (LLMs) on open-ended generation tasks. However, static rubric criteria may become saturated or remain beyond the policy's exploration, rendering the reward signals insufficiently discriminative. While recent efforts seek to guide sampling exploration or evolve rubrics, they fail to explicitly factor in how rubric difficulty should track the evolving policy's capabilities, resulting in a persistent mismatch between the optimization objective and the policy's current capability frontier. In this paper, we propose \textbf{Co}-Evolving Policy and Rubric via \textbf{C}apability \textbf{A}wareness (\textbf{CoCA}), a closed-loop framework that jointly adapts the rubric and policy through criterion-level diagnosis. Specifically, CoCA contrasts query-only and rubric-guided rollouts to characterize each criterion relative to the current policy, recalibrates criteria with mis
benchmark - arxiv:2610.38547 · cs.CVDetecting Fine-Grained Differences between Image DatasetsLeander Girrbach, Alexander von Recum, Kirill Bykov, Zeynep Akata
Detecting whether and in particular how two data distributions differ is a fundamental statistical problem. For example, we might be interested in which concepts (expressed in natural language) differ between two classes, the outputs of different generative models, or the same model with different prompt specifications. When working with large and complex image datasets, we intuitively think that first selecting concepts from a broad list using fine-grained patch embeddings, and then optimizing the concepts to better fit the data is an effective workflow. However, prior work instead uses MLLMs to propose concepts and image-level embeddings to select them. Therefore, we systematically compare these techniques and find that selecting concepts from broad lists can replace MLLM-proposals at obviously lower cost. MLLM inference can then be spent on optimizing selected concepts instead to achieve substantial improvements. Additionally, by designing a new benchmark that focuses on low-prevale
benchmark - arxiv:2610.38546 · cs.CLImplicit Minor Detection from Multi-Turn Dialogue in High-Risk AI Interactions: A Constrained Self-Evolving LLM HarnessXiaohan Zhang, Ying Zhao, Ruilin Guo, Xuanyuan Yu +4
In high-risk text interactions, a system may need to estimate whether the current speaker is likely to be a minor before applying age-conditioned safeguards. Many text-based dialogue systems lack strong identity signals such as account information, verified identity, facial data, or device data, and must rely on weak cues that emerge across multiple turns. We formulate implicit minor detection under weak identity information as a high-risk user-attribute inference task and propose a constrained framework comprising a large language model, a Harness, and a Maker-Checker protocol. The framework controls when inference is invoked, requires structured dialogue-linked evidence, and subjects validation-selected Harness revisions to version-level review. On a controlled Chinese benchmark comprising 8,328 multi-turn dialogues, our method improves average precision and minor-class F1 by 7.60 and 6.37 percentage points, respectively, over direct inference using the same LLM family.
self-evolvingbenchmark - arxiv:2610.38544 · cs.AIAn Alignment Benchmark for Enterprise Agents on Era by EonHai Sadon, Benjamin Gruenbaum, Doron Porat, Assaf Natanzon +3
AI agents are widely used in enterprises to access business systems and act on them in pursuit of a goal. However, fulfilling a user's request can cause harm. Era by Eon is a product that creates a digital twin of a company by simulating its business systems and data. Its simulators implement the programming interfaces of real products. Our benchmark uses ten of these systems, including Salesforce, Zendesk, Jira and Slack. We tested whether agents would carry out harmful business requests or trust misleading information in company records. We ran 19 models on 41 tasks three times, for 2,337 runs. Of the 2,304 runs that produced an answer, 918 (40%) contained a violation. The most common were false claims about actions or records (290 runs), following planted instructions (266) and disclosing sensitive data to the wrong audience (249). Every attempted change was recorded and blocked before it took effect. Misleading Slack messages caused 16 models to report an incorrect ticket status in
ai agentbenchmark - arxiv:2610.38536 · cs.ROInfluence of Agent Number and Behavior on Users' Psychological Responses in VR Evacuation TrainingNaoki Matsumura, Kensuke Sugihara, Tomoko Yonezawa
This study aims to build effective drill software by combining a virtual simulation of a disaster with multiple agents capable of expressing emotions. We evaluated how the number of agents and their emotional and behavioral actions during evacuation influence users in a virtual space. The results showed that users tended to act more confidently when many agents evacuated with them, and that the evacuation route chosen by the agents influenced the users' decision-making. These findings indicate that users' evacuation behavior is shaped not only by the environment itself but also by the collective behavior of surrounding agents, underscoring the need to consider how agent behavior affects users' psychological states during training.
agent - arxiv:2610.38532 · cs.AIFrom Leniency to Suspicion: How an Open Post-Training Recipe Changes Whom LLM Fact-Checkers BelieveNouar AlDahoul, Yasir Zaki
LLMs increasingly fact-check online content with the poster's identity in view. However, it remains unclear which post-training changes how much a model's truth judgement depends on the poster's social class, and how such a change happens inside the model. This paper investigates job-based doubt, i.e., judging a claim false more often when the poster has a lower-status job, using fact-checked claims with randomised poster profiles in base and post-trained checkpoints of six model families. Our results reveal that: 1) post-training creates such doubt only with the open Tülu-3 recipe (OLMo-2 32B, Tülu-3 70B), not with Meta's, Alibaba's or Mistral's; the effect is under one percentage point but statistically reliable, exceeding training drift and replicating on independent claim sets; 2) in OLMo-2 32B, the model does not start noticing the poster's job more, but starts interpreting it differently: a lower-status job pushes the base model toward "true" and the post-trained model toward "fa
post-training - arxiv:2610.38530 · cs.CVFrom the Superderivative to a Relative-Gradient Edge OperatorAlejandro Garnung Menéndez
First-derivative edge detectors such as Sobel, Prewitt, Roberts and Scharr measure absolute grey-level differences. We revisit the log-gradient magnitude, a relative-gradient edge operator, and motivate it by analogy with the superderivative, a multiplicative derivative in which the independent variable is perturbed multiplicatively rather than additively. In one and two dimensions the logarithm of the superderivative is a radial logarithmic derivative; its local, origin-free, isotropic analogue is the magnitude of the logarithmic gradient. The resulting operator measures relative rather than absolute change, giving an edge-strength map that is invariant under global intensity rescaling. We present the theory, show how the operator can be implemented in two steps---a classical gradient kernel plus a pointwise division---and compare it with standard first-derivative filters on synthetic images with ground-truth edges and on the official BSDS500 benchmark. We also study a parametric fami
benchmark - arxiv:2610.38508 · cs.ROSaiVLA-JEPA: Evaluating Future Latent Prediction in JEPA-Conditioned VLA ControlRufeng Xiao, Ruihao Xiao, Wei Chen, Xinhai Sun
Joint-embedding predictive architectures (JEPAs) learn to predict representations rather than pixels, offering a compact way to model how visual scenes evolve. We ask whether this predictive objective improves closed-loop control when a vision-language-action (VLA) policy already uses V-JEPA features. We call our JEPA-conditioned flow policy family SaiVLA-JEPA. It combines a vision-language model with a diffusion transformer action expert and receives features from a frozen V-JEPA encoder. Two lightweight auxiliary heads predict future representations during training and are omitted at inference, leaving the action sampler unchanged. We compare policies without JEPA features, with current JEPA conditioning, and with additional future supervision on 10,030 paired LIBERO-Plus episodes. Current JEPA conditioning reaches about 82.5% success versus 71.6% without JEPA, while future supervision provides no detectable aggregate gain at the evaluated training budget. We separately evaluate SaiV
vision-language-actionvlaliberov-jepa - arxiv:2610.38502 · cs.ROOmniQuad: a Wheeled-Legged Hybrid Robot with Omnidirectional WheelsFrancesco Iotti, Alok Ranjan, Franco Angelini, Manolo Garabini
Combining wheeled and legged locomotion within a single robotic platform offers the potential to harness the advantages of both systems. The development of such hybrid systems remains an active area of research. While legged robots exhibit remarkable adaptability for traversing uneven and rugged terrains, their relatively low speed limits their practicality in indoor environments. Conversely, wheeled robots demonstrate higher efficiency on flat surfaces but often encounter difficulties when navigating obstacles such as steps. This paper introduces OmniQuad, a novel hybrid robot that integrates mecanum wheels with legged locomotion to exploit the benefits of both modalities. Mecanum wheels provide omnidirectional mobility, enhancing the robot's manoeuvrability by enabling agile movement in confined spaces and the ability to maintain a consistent orientation during task execution. In this work, we first present the design of a custom wheeled-foot featuring mecanum wheels. Subsequently, w
legged locomotion - arxiv:2610.38499 · cs.CVResidual Motion Splatting for Continuous Space-Time Video Super-ResolutionVincent Herfeld, Eva Coupet{é}, Julien Mairal, Jean Ponce
Continuous space-time video super-resolution produces a high-resolution, highframe-rate video from a low-resolution, low-frame-rate one at arbitrary spatial scales and temporal instants. Classical spatial multi-frame super-resolution has shown that detail beyond a single frame's sampling limit can be recovered by aligning complementary observations with sub-pixel precision and accumulating them on a finer grid. Existing learning-based methods largely depart from this principle: some align and fuse frames only at low resolution before regressing a continuous high-resolution representation, while others interpolate each frame to the output resolution before splatting. In both cases, the per-frame samples are resampled before fusion, so their sub-pixel positions are lost before frames are combined. Restoring this principle in a learned continuous model raises two challenges: estimating sub-pixel motion at unobserved times, and handling holes and occlusions when splatting a coarse grid ont
benchmark - arxiv:2610.38497 · cs.CVVision-Based Density Estimation for Adaptive Traffic ControlMohammad Usaid, Deepak Silaych, Hrithik Mhatre, Archak Mittal
Traffic congestion is becoming a critical challenge in rapidly growing urban centers, particularly in developing nations where rapid urbanization and inadequate transport infrastructure place increasing strain on transport networks. In highly heterogeneous and chaotic traffic flows, camera-based traffic-state estimation is limited by severe vehicle occlusion and overlap. In this paper, we propose a framework to optimize traffic flow directly from stochastic visual observations. The architecture is structurally decoupled into three primary components: (a) vehicle density estimation, (b) predictive dynamics modeling, and (c) policy optimization. The proposed methodology introduces a convolutional neural network (CNN) that learns the mapping from camera feeds to spatial density maps, utilizing an adaptive k-nearest-neighbor (k-NN) Gaussian kernel to handle varying congestion levels. This feeds into an ST3DNet predictive module for spatiotemporal forecasting, followed by a hierarchical Gra
multi-agent - arxiv:2610.38496 · cs.LGAccelerating Nonlinear MPC by Behavior Cloning with Separated Timescales: Application to Nuclear Reactor Load-FollowingPerceval Beja-Battais, Alain Grosset{ê}te, Nicolas Vayatis
Behavior Cloning (BC) of expert controllers for multiscale dynamical systems is challenging: state variables evolving at widely separated timescales induce highly nonlinear couplings that a shared latent space may fail to adequately represent, degrading closed-loop generalization. We propose Structured Representation Learning for Behavior Cloning (SRLBC), an architecture designed to leverage the physical knowledge of multiscale systems in which slow state variables, fast (stiff) state variables, and predictable exogenous inputs evolve at different timescales. On 500 long-horizon roll-outs of a pressurized water reactor (PWR) load-following benchmark -a high-dimensional, stiff industrial system with challenging safety constraints -and across 20 training seeds, separated embeddings systematically outperform shared embedding baselines on cumulative cost, imitation accuracy and constraint satisfaction. The resulting standalone policies approach NMPC closed-loop performance at a computation
benchmark - arxiv:2610.38491 · physics.app-phTwo-Dimensional Magnetic Materials from Ferromagnetism to Altermagnetism: Fabrication, Symmetry, Evidence, and Device ReadinessGuojian Zhao, Ziang Meng, Li Liu, Zhiyuan Duan +8
Atomically thin magnets are governed by composition, thickness, stacking, strain, defects, substrates, ferroic order, and interfaces. This Review organizes two-dimensional (2D) magnetism using net magnetization and momentum-space spin splitting or spin-polarization patterns as primary coordinates, supplemented by sublattice structure and the symmetry relations between opposite-spin sublattices. Ferromagnets generally carry a net moment; ferrimagnets can be fully compensated. Conventional antiferromagnets may retain spin-degenerate bands, while altermagnets permit nonrelativistic alternating spin splitting under dimension-specific symmetry constraints. M1-M4 reporting classes introduced here distinguish layer-resolved atomically thin forms, non-layer-resolved thin flakes or films, bulk benchmarks, and theoretical candidates or design routes; the evidence category and ordering-temperature provenance are reported independently. Current experimental support for altermagnetism comes mainly
benchmark - arxiv:2610.38482 · cs.CVGeometry-Aware Trajectory State Modeling for Few-Shot Action RecognitionZhiqi Miao, Marko Haralović, Stanislav Levendeev, Sarvesh Angadi +4
Few-shot action recognition (FSAR) requires representations that expose discriminative motion dynamics from only a few labeled videos. Recent trajectory-based methods leverage long-term point tracks for explicit temporal correspondence, but often treat trajectories as carriers of sampled visual features or summarize their motion with compact 2D displacement-based descriptors. We instead view an action as the ordered evolution of 2.5D trajectory states and their local geometric interactions. Based on this view, we introduce GeoTRAM, a geometry-aware trajectory state modeling framework for FSAR. GeoTRAM lifts sparse tracks into 2.5D trajectories, where StateMotion models ordered intra-trajectory state evolution and RelMotion captures state-conditioned local interactions through time-varying relative geometry and motion. Quality-filtered spectral sampling provides more reliable trajectory initialization; complementary semantic, motion, and depth cues are separately refined and preserved f
benchmark - arxiv:2610.38468 · cs.MACinematize from Unreal: Agentic Choreography for Controllable World SimulationZifan Song, Caihua Shan, Kaitao Song, Yifan Yang +2
Controllable world simulation requires more than visually plausible frames: scenes must retain persistent state, coordinated actions, and editable temporal structure. Generative world models offer flexible visual synthesis, but their outputs do not always expose the objects and transitions needed for precise revision. Graphics engines provide executable state and direct control, yet translating language into valid operations requires specialized knowledge of assets, geometry, and interdependent APIs. We present CineU (Cinematize from Unreal), an agentic system for language-driven world orchestration and cinematic choreography within Unreal Engine. From a code-based world modeling perspective, an Unreal project serves as a persistent, inspectable world representation, while agent-generated code specifies operations whose effects can be executed and verified. CineU combines hierarchical orchestration, cascaded asset retrieval and geometry-constrained placement, capability-aligned code sy
world modelpersistent stateagentic - arxiv:2610.38458 · cs.AITracekit Box: Running AI Agents in Micro-VMs Where Every Way Out Is Decided and SignedBravish Ghosh
AI agents now install packages, call model APIs with their operators' keys and change code in repositories that deploy to production. The network, credentials and write access they need are also how they can leak data, steal a key or change a system nobody approved. Sandboxes contain an agent but leave little evidence of what crossed their boundary; tracing records what happened but contains nothing. We present Tracekit Box, an open-source system (MIT, v0) that does both. Each job runs in its own Firecracker micro-VM with no network card. The only ways out are doors on the host, served over vsock: commands, outbound connections, secrets, model calls and the workspace patch. At every door a Tracekit signer the agent cannot reach decides what passes and signs it into one run per box. Names are resolved on the host; private and cloud-metadata addresses are always refused; secrets never enter the VM, which sees placeholders the host proxy swaps for real values only in one header for bound
agentai agent - arxiv:2610.38452 · cs.CLWhen Support Crosses the Line: Evaluating Ethical Boundaries in LLM-Based Mental Health Support in the Chinese ContextBichen Wang, Yufeng Jin, Yixin Sun, Yanyan Zhao +1
Large language models (LLMs) are increasingly explored for mental health support, yet existing benchmarks primarily focus on supportive capabilities rather than whether models consistently adhere to professional ethical boundaries. This issue is particularly important in multi-turn interactions, where users may introduce new information or exert emotional pressure over time. We introduce MindEth-Bench, a benchmark for evaluating ethical boundary adherence in multi-turn mental health interactions. Grounded in professional ethical standards for mental health services in China, MindEth-Bench covers multiple dimensions of ethical conduct and uses Monte Carlo Tree Search (MCTS) to systematically explore interaction trajectories that stress-test model boundaries. Experiments across nine LLMs show that models that initially respond appropriately can shift toward ethical violations over extended interactions, with substantial variation across models and ethical dimensions. These results show t
benchmark - arxiv:2610.38450 · cs.CLCoSTA: Co-Evolving Self-Play Timeline Agents for Complex Temporal Knowledge Graph Question AnsweringHaikun Liu, Zhichao Yan, Shaoru Guo, Jiawei Gu +3
Complex Temporal Knowledge Graph Question Answering (TKGQA) involves more than retrieving facts: the timestamp or validity interval of one fact can constrain the search for another, while multiple intervals may jointly determine the answer. Existing single-agent TKGQA methods typically investigate these facts within a single reasoning context, where evidence from different event histories can become interleaved and distractors can misdirect subsequent retrieval. Distributing investigations across agents can separate event-specific evidence into distinct contexts, but effective collaboration must still be learned because question decomposition, subtask investigation, and evidence synthesis are interdependent. To address these limitations, we propose \textbf{Co}-Evolving \textbf{S}elf-Play \textbf{T}imeline \textbf{A}gents (CoSTA), which organizes facts into navigable timelines, delegates investigations to workers, and uses a coordinator to synthesize their original timestamped observati
knowledge graphllm agentself-playbenchmark - arxiv:2610.38448 · cs.AIBenchmarking decision models against generative large language models on medical multiple-choice questionsOscar Freyer, Daniel Glauert, Moritz Hofmann, Max Ostermann +2
Medical workflows often require selecting among predefined options. We compared 13 decision models with 15 generative large language model configurations, including different reasoning settings, using 1,273 four-option MedQA questions under text-only conditions. Accuracy ranged from 67.71% to 89.08% for decision models and from 78.32% to 96.86% for generative configurations. Among the tested configurations, Mercury Decide had the lowest mean final-attempt cost (USD 0.0069 per 1,000 questions), while Clef Omni had the shortest median request duration (0.291 seconds). Decision-model probabilities distinguished correct from incorrect answers, but error rates among answers assigned probabilities of at least 0.90 ranged from 1.71% to 12.77%. Answer agreement across three runs ranged from 92.30% to 100%, while probability repeatability varied. Decision models offered resource advantages alongside lower peak accuracy; selecting models for medical workflows requires considering probability qua
benchmark - arxiv:2610.38443 · cs.ROSkeleton-Guided Geometry-Aware Auto-Labeling for Robotic Grasping: From 4-DoF Learning to 6-DoF ExecutionAli Sabzejou, Farbod Azimmohseni, Mehdi Tale Masouleh, Ahmad Kalhor
Robotic grasping relies on large manually annotated datasets, which are costly to create. We introduce a straight-skeleton-based algorithm for generating two-finger grasp keypoints and identifying regions suitable for stable, feasible grasps. Our automated pipeline produces consistent grasp annotations without human intervention. We also propose the Skeleton-based Generative Grasp Convolutional Neural Network (SkelGG-CNN), a 4-DoF architecture that learns geometric grasp features from skeleton-based guidance. To extend auto-labeling to 6-DoF grasping, we develop a generalized quasi-3D skeletonization algorithm that operates directly on point clouds. Using automatically generated 4-DoF labels, SkelGG-CNN achieves accuracies of 95.42%, 96.55%, and 97.1% on Jacquard V1, Jacquard V2, and Cornell, respectively. On the GraspNet novel split, 6-DoF auto-labeling achieves an average precision of 18.33, comparable to state-of-the-art model-based approaches. In simulation, the 6-DoF method achiev
grasp - arxiv:2610.38439 · cs.AIPredict the Reward, Not the World: World Reward Models for ML AgentsYuxiao Lee, Zherui Yang, Fan Liu, Hao Liu
Search-based machine-learning (ML) agents improve a solution by writing candidate programs, executing them, and keeping the best. Execution gives the search reliable feedback, but a single run can take hours, and most runs end in programs the agent discards. Recent agents try to save these runs by letting a large language model act as a world model that predicts how each program would perform. However, a wrong prediction feeds the search false information. We argue that an agent does not need to predict the outcome of execution, only the decision that execution would support. We propose World Reward Models (WRMs), which learn from real execution results to choose directly which candidate will perform better, without predicting any intermediate outcome. We prove that a selector which predicts execution from the same inputs cannot outperform the best direct selector, that predicted outcomes add an error which can make a likelihood-trained world model prefer the worse program on every ins
world modelagent - arxiv:2610.38438 · cs.AIRoleFUZZER: User-Behaviour Fuzzing for Role-Specification Adherence in Large Language ModelsKshitish Ghate, Michael Saxon, Tadayoshi Kohno, Aylin Caliskan
Large language models (LLMs) are often conditioned into specific roles such as tutors, interviewers, or agents acting for an organisation, through system prompts that specify how they should behave. The extent to which they can actually adhere to these role specifications under multi-turn interactions with realistic and potentially problematic users is understudied. To address this, we propose user behaviour fuzzing, a guided search over plausible user behaviours that can lead a model to violate its role specification, and introduce RoleFUZZER, a framework that adapts coverage-guided fuzzing from software testing to operationalise the search. RoleFUZZER constructs scenarios guided by a user-behaviour coverage grid that crosses a grounded taxonomy of user interaction types and the role specification of the model under test i.e. what is under pressure. With RoleFUZZER, we evaluate LLMs across five role specifications, including tutor, financial advisor, AI interviewer, antagonistic rolep
agent - arxiv:2610.38437 · cs.LGCapacity of the Hopfield-Hebbian dynamics with correlated patternsMihailo Stojnic
Allowing for small fraction of errors in retrieved stored binary patterns, Hopfield-Hebbian neural network (NN) associative memory dynamics [14] achieves linearly scaled capacity. The famous proportionality constant $α_c\approx 0.14$ is obtained in a large dimensional scenario as the ratio of the number of reliably stored patterns $m$ and the patterns' length $n$ under the assumption that uniformly randomly chosen binary patterns are \emph{independent} among themselves. While such an idealized scenario is now well understood, more practical scenarios with correlated/noisy datasets remaining challenging even for classical statistical mechanics replica approaches. We consider precisely such practical settings where each \emph{archetype} pattern is replaced with $M$ correlated \emph{examples} with parameter $r$ governing the quality of correlation. We develop a powerful Random Duality Theory (RDT) based framework to analyze correlated data and already on the first level of RDT lifting obt
memory - arxiv:2610.38436 · cs.LGContinual Adaptation Across Modalities: Learning and Forgetting in Unified Multimodal ModelsXinrui Wang, Zhihao Chen, Jia Tang, Alexandra Gomez-Villa +3
Unified multimodal models (UMMs), which can process and generate diverse modalities within a single framework, have recently emerged as a powerful paradigm. However, when these models undergo continual learning on multiple tasks across different modalities, their learning and forgetting dynamics remain poorly understood. In this paper, we systematically investigate how architectural design and adaptation method shape intra-modal learning and cross-modal interference between the generation and understanding sides. Specifically, we report three findings. First, the severity of cross-modal interference decreases with the degree of modality-specific parameter isolation. Second, a within-model asymmetry: the generation side of a UMM is markedly more fragile than its understanding side. Third, what post-training adds is the first to be forgotten: purely post-training capabilities such as personalized generation collapse by 84\% to 90\%, while pretraining capabilities remain stable. Together,
post-training - arxiv:2610.38433 · cs.LGDoes Chain-of-Thought Reliance Follow a Scaling Law?Larger Models Follow Their Steps More, Independent of Post-TrainingEsther Xin
Monitoring a model's chain of thought (CoT) is useful for safety only if the final answer depends on the written steps. Prior measurements use natural tasks where true reasoning is unknown. We measure it against an exact ground truth on a synthetic box-swap task where every intermediate state is known. We replace one step inside a model's own chain with a wrong but valid state and check whether the answer follows the edit. We evaluate 25 open-weight models--Qwen3 (0.6B-14B, base and post-trained), Gemma 3 (1B-12B), Llama 3 (1B-8B), and 8 OLMo 2 checkpoints--on 7,200 items each. We find five results: (1) Reliance rises with size and saturates. In Qwen3, answer reliance rises from 16-18% at 0.6B to 97-100% at 14B; a logistic in log N fits well (R^2 = 0.89), and Gemma 3 and Llama 3 follow this trend. (2) Capability is the main driver. Across all models, reliance and step accuracy correlate at r = 0.99; equally capable models are equally reliant, regardless of family. (3) Size still matter
post-training - arxiv:2610.38426 · cs.AIThe Code Behind the Layout: AI Agents for Analog Layout GeneratorsMinxing Chu, Anhang Li, Martins Atilla, Ruichen Qi +1
Analog layout remains largely manual and hard to automate. Generator frameworks such as gLayout express layouts as programs, producing layouts from parameterized, design-rule-aware generators; however, writing such generators still takes expert effort. Large language models (LLMs) could now write generator code, and frontier models can even draw layout geometry from design intent. This raises a practical question: if a model can draw the layout itself, what does a generator framework still contribute, and is applying these models to a generator framework better than having them draw a layout from scratch? Through controlled experiments on gLayout in SKY130, we find that the main obstacle to drawing a layout directly is its time and token cost, and that the main advantage of a model building on a generator framework is reusability. Its verified primitives keep each model edit small enough to check, and its parameters allow a single edit to be evaluated across an entire family of layouts
ai agent - arxiv:2610.38425 · cs.CVIntReal: A Real-World Dataset and Benchmark for Multimodal Interleaved GenerationShaohan Yu, Jie Qin, Wende Tan, Haiwen Diao +3
Humans routinely consume multimodal content, where textual and visual information is naturally interleaved. Although Unified Multimodal Models (UMMs) can jointly generate text and images, multimodal interleaved generation remains largely underexplored. Progress in this area is hindered by two key limitations: (1) the lack of well-defined interleaved generation tasks and corresponding training data, and (2) the absence of fine-grained evaluation protocols that can reliably diagnose model performance across different capability dimensions. To this end, we present IntReal, comprising two complementary components: IntReal-100K, a real-world dataset of multimodal interleaved samples spanning four major categories, and the IntReal Evaluation Suite, which evaluates model performance across seven complementary dimensions using specialized expert models, enabling coarse-to-fine diagnosis. Extensive experiments validate our evaluation suite through low inter-metric redundancy (max pairwise absol
post-trainingbenchmarkevaluatorevaluation protocol - arxiv:2610.38411 · cs.AIGovernance-Aware Federated Learning for Multi-Agent AI Systems: A Policy-Driven Framework for Trustworthy Collaborative AIKiranmayee Janardhan, Karthick S, Madhusmita Panda, Bharathi Vokkaliga Ganesh +1
Federated Learning (FL) enables collaborative model training without exchanging raw data, but conventional FL provides limited runtime governance, evidence verification, fairness assessment, explainability monitoring, and auditability. This paper proposes Governance-Aware Federated Learning (GAFL), which combines independent governance-evidence verification, temporal trust, fairness and explanation assessment, anomaly detection, policy enforcement, normalized aggregation, and tamper-evident audit logging. Client-reported governance values are authenticated, bound to their submitted model updates, and independently recomputed before aggregation. A cross-silo prototype with 10 participants was evaluated on the Adult Income dataset under honest, benign non-IID, biased-data, malicious-update, non-compliant, falsified-evidence, and mixed-threat scenarios. Four methods were compared across five random seeds: FedAvg, FedAvg with anomaly filtering, GAFL without independent verification, and co
multi-agent - arxiv:2610.38407 · cs.CLHolistic Diversity Management in LLM-Guided Evolutionary SearchSeockbean Song, Youngsik Yoon, Sungjae Lee, Siwei Wang +2
LLM-guided evolutionary search iteratively selects existing programs, uses LLMs as mutation operators to generate a new program, and evaluates it to guide subsequent search. To preserve diversity and reduce the risk of premature convergence, it often adopts the island model, in which groups of programs, called islands, evolve separately. However, this structure alone does not guarantee distinct search directions, as islands may start from similar programs or converge toward similar solutions during evolution. This suggests that diversity must be considered throughout the search process, from initializing islands with distinct programs to developing island trajectories and sharing discoveries across islands. To this end, we propose HoliDivEvolve , a holistic approach to establishing and effectively utilizing island diversity throughout evolutionary search. Specifically, we introduce diverse initialization to establish different initial programs, trajectory adaptive island selection to a
benchmark - arxiv:2610.38405 · cs.ROSpaceAct: Benchmarking 3D Spatial and Temporal Reasoning via Closed-Loop Action and Active PerceptionChenglong Shi, Zekai Yuan, Yue Song, Jon Pu +1
To solve problems in 3D space rather than merely describe images, multimodal large language models (MLLMs) need 3D spatiotemporal reasoning: reconstructing 3D geometry from occluded 2D views, anticipating physical dynamics, and executing precise, closed-loop actions. Prior studies test passive VQA on static images or symbolic planning, requiring no execution or supplying privileged coordinates while ignoring collision, statics, and timing. We introduce SpaceAct, an interactive benchmark evaluating 3D spatiotemporal reasoning via closed-loop physical action and active perception, with SAPA, a Spatiotemporal Active-Probing Agent framework. Built on interlocking bricks and mechanical puzzles, it has 108 verified instances across 36 task families in six categories, without privileged coordinates, and diagnostic modes (+plan, +state, no-camera, blind, L7 levers). In brick-bridging or aperture-threading, models rotate the camera to resolve occlusion, read in-scene rulers, output 6-DoF poses
agentagent frameworkbenchmark - arxiv:2610.38400 · cs.ROWhen "No Worries" Went Wrong: HRI Lessons for Foundation-Model RobotsEliahu Khalastchi
When a person declined our robot's invitation to play, Alice turned her shoulder away and said "No worries, it's all good." Only after a brief pause did she drive off. The words and gestures were plausible but their order changed the message. In another episode, Alice informed a host that his visitor had arrived---while the visitor stood right beside him. Both came from a three-year ROS 2 HRI consortium that integrated large language and vision models as they emerged. Today's foundation models give robots broader capabilities to perceive their surroundings, reason about situations and generate behavior. Yet capability alone does not ensure that a changed situation shapes the action that reaches a person. Evidence may never be acquired, or a purpose or constraint may fail to survive planning and execution. A system may even report Success when an action no longer serves its intended purpose. Drawing on the consortium's engineering experience, we develop five design lessons about attenti
embodied - arxiv:2610.38398 · cs.AIAgent Transfer Protocol: Server-Mediated Agent-to-Agent Communication Architecture for the Agentic WebYuqi Qiu, Zuyao Xu, Lu Sun, Xiang Li +4
The Agentic Web needs to support continuous collaboration among agents from different organizations. These collaborative relationships may span multiple tasks, network connections and execution instances. Such interactions require stable communication entries, message reception when instances are offline, and source verification after delayed reading. How to organize these responsibilities so that cross domain communication does not depend on specific execution instances being continuously online is a key issue for building long term agent interaction infrastructure. We propose ATP (Agent Transfer Protocol), a cross organization agent communication protocol adopting a domain mediated architecture. In this architecture, administrative domains maintain stable addresses and receiving entries for logical agents. The domain servers of both parties continuously undertake cross domain transmission. The receiving domain verifies and saves messages for execution instances to read later. The sen
agentllm agentagentic - arxiv:2610.38392 · cs.AIMasked Erosion: How Acceptance Gates Lose Redundant Safeguards in Self-Improving AgentsHau-Sing So, Daixian Liu, Weicai Long
Self-improving LLM agents edit their own harness, the prompts and rules around a fixed model, and keep an edit only if an acceptance gate scores it higher on a batch of tasks. Such a gate sees what the harness does under the current model, not what it would still prevent if one safeguard failed. When safeguards overlap, deleting one is masked: the others still block the harm, so the gate decides on task success alone and redundancy erodes unseen. We measure harm for every subset of four written safeguards in two agent environments and run 411 improvement loops under ten gate rules, mostly as controlled deletion experiments, plus three runs of the GEPA optimizer. A deletion behind another working safeguard adds almost no observable harm. A paired significance test blocks deletions that pass by chance but not deletions that genuinely raise task success, and a larger evaluation batch lets it delete cheaper safeguards: at the same cost, 20-task batches kept every harness protected while 40
agentllm agentself-improving - arxiv:2610.38391 · cs.AISCOPE: Selective Critical-Step On-Policy Self-Distillation for Agentic Reinforcement LearningQirui Liu, Wei Cao, Yu Mi, Yichen Sun +4
Outcome-based reinforcement learning provides coarse supervision for long-horizon LLM agents: failed trajectories may contain correct steps, while successful trajectories can contain errors. Trajectory-level skills can also shift teacher preferences at unrelated steps, introducing off-target distillation signals. We introduce SCOPE, a framework for Selective Critical-Step On-Policy Self-Distillation. After a supervised warm-up, the policy analyzes its own trajectories, generates step-specific hindsight skills, and distills their guidance only at selected error steps. A bounded signed teacher-student coefficient stabilizes auxiliary updates while preserving positive reinforcement and negative correction. A success-relative pairwise length reward further encourages efficient task completion. Experiments on ALFWorld and WebShop demonstrate consistent improvements across multiple model backbones. With Qwen2.5-3B, SCOPE achieves 97.7% success on ALFWorld and 85.2% on WebShop, outperforming
llm agentagentic - arxiv:2610.38389 · cs.CVVisually Irrelevant Hallucinations in Medical Vision-Language ModelsKe Zou, Nan Zhou, Tao Yang, Meng Liu +8
Medical vision-language models (VLMs) show strong performance in radiology report generation (RRG) and medical visual question answering (Med-VQA), producing free-form clinical descriptions and answers from medical images. However, they may generate clinically meaningful statements that lack visual evidence or contradict image content. We refer to these errors as visually irrelevant hallucinations and investigate them at the phrase level. We propose CLEAR (Counterfactual Localized Evaluation of Alignment and Response), a model-agnostic post-hoc framework for phrase-level hallucination risk estimation. CLEAR decomposes generated responses into visually assessable clinical phrases. For each phrase, phrase-conditioned grounding identifies a candidate visual region, which is selectively perturbed by a localized counterfactual intervention. CLEAR characterizes the intervention using two complementary auditing cues: response cues, capturing semantic and clinically meaningful changes in the g
benchmark - arxiv:2610.38388 · cs.ROELEVATE: Agentic Controller Acquisition from Task Failure for Quadrupedal ManipulationShiqi Liu, Radu Corcodel, Ding Zhao
Quadrupedal manipulators can extend their workspace through whole-body interaction, but task-planning agents' capabilities remain limited by their available controllers. We present ELEVATE, a framework that uses persistent failures during task-program refinement to identify and acquire missing whole-body control capabilities. When these failures suggest a missing physical capability, a coding agent formulates and iteratively refines a reinforcement-learning problem, specifying behavior-specific rewards, curricula, and termination conditions. Validated controllers are added to the capability library and composed with existing controllers to retry the task. Starting from a single pretrained loco-manipulation controller, ELEVATE autonomously acquires controllers for wall-supported elevation, supported manipulation, and wall descent across a sequence of seven simulation tasks. These controllers are reused in subsequent reaching and object-manipulation tasks. On the five tasks beyond the in
manipulationmanipulatorquadrupedwhole-body controlagentagentic - arxiv:2610.38385 · cs.LGRevisiting Temporal Link Prediction through Structural ProximityYuzhang Fei, Yanping Zheng, Lu Yi, Zhewei Wei
Existing temporal link prediction methods typically rely on learned temporal representations to capture evolving interaction patterns. In this paper, we show that a classical structural proximity measure can match or outperform these methods without training or learned temporal sequence modeling. We revisit Personalized PageRank (PPR) on historical interaction graphs, using simple sliding-window retention or exponential time decay. Surprisingly, across 14 datasets, Time-decay PPR ranks first on nine and second on four among the compared methods, including datasets with few repeated edges. To understand this result, we examine the historical paths connecting source nodes to previously unseen destinations. We find that such paths contain useful predictive information, and that PPR outperforms a strong neural baseline on queries where historical structure clearly distinguishes the true destination from negative candidates. These results show that strong performance on current benchmarks d
benchmark - arxiv:2610.38382 · cs.AICompute Has Shape: Self-Morphing the Internal Computation of Frozen LLMsJunxi Wu, Yiwen Guo, Shu-Tao Xia
Test-time scaling improves large language models through additional computation, but how to adapt the topology of internal computation to individual inputs remains underexplored. Existing approaches largely fix recurrent or parallel architectures during training, or reconfigure frozen models along a single structural axis. We introduce SelfMorph, which makes computational shape an input-dependent inference-time decision. It unifies the original forward path (BASE), serial reuse of a layer span (REPEAT), and local branching and merging of perturbed hidden states (BRANCH) in a unified space parameterized by type, location, and intensity. Since they offer complementary benefits rather than uniform dominance, we use a lightweight selector that reads the backbone's depth-wise hidden state trajectory on the prompt to choose a morphology before response generation. We train the selector by distilling utility-based distribution from offline morphing that jointly account for task performance an
benchmark - arxiv:2610.38366 · cs.CLTCC-RAG: Temporal Causal Chain-Guided RAG for Temporal Knowledge Graph Question AnsweringXinxun Zhang, Yuxiang Wang, Shanshan Feng, Jun Zhang +5
Recent advances in large language models (LLMs) have spurred growing interest in temporal knowledge graph question answering (TKGQA). Most existing methods rely on semantic retrieval augmented generation (RAG), which suffers from fundamental limitations in complex temporal reasoning. On the one hand, semantic-based RAG methods struggle to capture the multi-hop causal knowledge required for complex temporal reasoning. On the other hand, these methods introduce excessive redundant knowledge, trapping LLMs in spurious reasoning chains that potentially result in erroneous answers. To address these problems, we propose a Temporal Causal Chain-guided RAG (TCC-RAG) to retrieve a concise yet sufficient set of causal knowledge to support complex temporal reasoning. Specifically, we reformulate the temporal knowledge graph into a multi-view causal cue graph, with causal cues inducing connections that explicitly encode temporal evolution and causal dynamics. Based on the causal cue graph, we cons
retrieval augmentedragknowledge graphbenchmark - arxiv:2610.38361 · cs.AIManifest-Assisted Retrieval in Artificial Intelligence Search: Evidence Gaps and a Security EvaluationGangzhen Qian, Kainan Zhou, Chuhong Xu, Hang Xiao +4
The external sources retained in our non-systematic audit did not isolate the incremental security or utility effect of consuming /llms.txt while ordinary retrieval remained fixed. The initial audit covered 39 sources through August 16, 2026, with later targeted updates through October 14. A purposive 15-tool study observed no automatic requests in 45 trials at one endpoint; a non-probability log study found that 97% of domains with detected files received no recorded requests in May 2026. We separate publication, fetch, model use, and action; define five operational threat paths and paired security-utility estimands; and exercise deterministic pre-model controls with 30 author-defined parser fixtures. We also execute a reduced pilot: 24 tasks across eight sites, five conditions, and 120 real model responses, with ordinary evidence held fixed. Frozen automatic scoring identifies attacker-target adoption in $21/24$ raw adversarial responses, with one further response unresolved; the adv
agent - arxiv:2610.38358 · cs.CVVolley3D: Unified 3D Full-Sequence Reconstruction from Monocular BroadcastsNoâm Boussouf, Adrien Maglo, Mohamed Chaouch, Hideo Saito
3D reconstruction from single-view sports broadcast is a pivotal capability for automated analytics and immersive AR/VR replays, yet notoriously ill-posed due to difficult-to-calibrate moving cameras and depth ambiguity. Unlike widely studied racket sports, team sports such as volleyball feature unpredictable multi-contact sequences and high altitude jumps. These dynamics categorically invalidate the standard planar projection assumptions relied upon by existing methods. To resolve these limitations without 3D training supervision, we propose a complete, physics-grounded pipeline tailored to monocular broadcast volleyball. First, we include a correction module to the perspective drift induced by ground-plane projection by coupling temporal jump spotting with Newtonian interpolation for 3D player placement. Second, for 3D ball trajectory estimation, we design a dedicated sub-trajectory segmentation architecture alongside a collision-matching algorithm. By anchoring physical equations to
benchmark - arxiv:2610.38357 · cs.ROCSO: Refining Robotic Policies via Skill Distribution Alignment and Skill-Grained OptimizationZhiyuan Xiang, Xiang Deng, Wanjie Tao, Xu Chen +2
Discretizing continuous actions into discrete skills using methods like VQ-VAE has emerged as a powerful paradigm for robotic manipulation. However, quantization errors yield a suboptimal distribution for the skill prior, degrading its performance. While reinforcement learning offers a path for refinement, its direct application faces unstable encoder updates and a granularity dilemma in importance sampling. To address these challenges, we introduce Cascaded Skills Optimization (CSO), a two-stage post-training framework. First, Rejection-Sampling Supervised Fine-tuning aligns the model's observation-to-skill mapping with the distribution of successful online trajectories via supervised fine-tuning. Second, with the observation encoder frozen, Skills Policy Optimization computes an independent, clipped importance ratio for each complete skill, enabling stable and efficient updates. With a lightweight QueST-style policy, CSO raises overall LIBERO success from 80.1% to 97.5%, improves per
manipulationliberorobotwinpost-training - arxiv:2610.38352 · cs.CVWeVisCut: The Whole Editing Room in One Agent From Narration to Long-Form VideoChenyu Mu, Wenming Weng, Qinrui Wang, Xingsong Ye +7
Producing long-form videos from narration requires interdependent decisions that remain coherent across shots. Existing agentic editors often use multi-stage workflows driven by general models, leaving the quality-cost trade-off of compact, task-adapted policies unclear. We address this gap with WeVisCut, a long-form script-to-video system that connects tool execution with policy learning. Its agent harness maintains shared production state so that tool feedback informs subsequent decisions and repairs. Because finished videos do not expose these decisions, our four-stage data pipeline combines editing inversion with verified replay to turn recovered editing intent into executable supervision. Applied to 3,750 finished videos, this pipeline yields over 130K training examples. We use this supervision to train a single 9B policy that coordinates production and recovers from failures within the same harness. Moreover, we introduce a 100-task benchmark and five metrics assessing editing qu
agentagenticbenchmark - arxiv:2610.38351 · cs.ROProprioceptive Finger Force Estimation for Reliable Box ManipulationJunha Cha, Chanwoo Kim, Kisang Park, Yeoncheol Jang +4
Reliable box manipulation requires successful placement without damage, which depends on maintaining sufficient contact force without excessive loading. We hypothesize that per-finger force feedback improves manipulation reliability relative to open-loop finger commands. To this end, we present Finger External-Force Estimation from Proprioceptive SignaLs (FEEL), an uncertainty-aware residual method that estimates per-finger external force magnitude from proprioceptive histories. Our method combines a calibrated virtual-spring estimate with a residual distribution predicted by a Mixture Density Network (MDN) and requires no dedicated contact sensors during deployment. The estimated force is used for continuous per-finger adaptation, while predictive uncertainty is evaluated as an empirical signal of ambiguity in proprioceptive force estimation. Across ten fingers and contact forces spanning 0-20 N, our method achieves a mean RMSE of 0.251 N, reducing the error by 92.8% relative to the v
manipulation - arxiv:2610.38349 · cs.CVAligning Misaligned Benchmarks for Single-View Novel View SynthesisYang-Che Tseng, Erqun Dong, Jeong-gi Kwak, Gopal Sharma +4
While single-view novel view synthesis methods have been dramatically enhanced recently, how to quantify their performance has yet to be standardized. In this work, we demonstrate that an important factor lies in how each method's own reference frame is aligned with the so-called ground-truth pose. Existing methods often rely on test-time optimization of target poses for this alignment, which inherently is ill-posed due to the complexity of the optimization landscape of camera poses vs render accuracy, leading to methods being evaluated at an arbitrary local optimum. Some recognize this limitation and thus rely on 3D reconstructions of entire scenes using training and test views to align poses, which may be infeasible. In this work, instead, we propose to use traditional 2D-to-2D and 2D-to-3D image matching techniques to align 3D reconstructions. We also evaluate only within co-visible regions, between different views, which is the only place where a `correct' ground truth can be defin
benchmark - arxiv:2610.38342 · cs.AIInvestor's Last Exam: Can AI Agents Do Equity Investment Without Seeing the Future?Huangfei Jiang, Anning Hu, Kai Sheng, Yuchen Cai +6
Financial agents must form forecasts and investment judgments from decision-time evidence, with predictions evaluated against subsequent outcomes. Existing benchmarks assess financial knowledge, research workflows, or forecasting, but evaluating the full research process requires a strictly frozen information set and professional evaluation criteria. We introduce Investor's Last Exam (ILE), an equity-research benchmark led by experienced buy-side investors and fund managers. Its three components are a versioned financial database spanning 3.43 million document versions and nine source categories, supplemented by structured market data, with validated timestamps governing retrieval and computation; a task taxonomy organized around professional analyst deliverables; and an expert-led framework linking scoring criteria to sources, calculations, and assumptions through financial dependency graphs. Evaluation spans 86 assignments and jointly assesses research quality, forecast accuracy, and
ai agentbenchmark - arxiv:2610.38336 · cs.AIStatutory Article Retrieval from Arabic Judicial Narratives: Family Law and Law of Evidence in Saudi ArabiaZarah Shibli, Areej Alhothali, Lulwah Alharigy
Reliable legal information retrieval must ground legal reasoning in authoritative statutory sources. In Arabic judicial settings, long factual narratives must be mapped to concise articles that express abstract legal rules, creating a narrative-to-legal-rule abstraction gap. Using a benchmark of 200 synthetic Saudi Arabic judicial cases and 381 articles from the Family Law and the Law of Evidence, we compare word-level and char-4 BM25, zero-shot dense encoders, per-law retrieval, dense-sparse reciprocal rank fusion (RRF), and a proposed Decompose-and-Refine (DaR) pipeline. In DaR, a large language model splits each narrative into disputed legal issues and generates, for each issue, three retrieval channels: a statute-style paraphrase, legal keywords, and a candidate taxonomy chapter. On the whole corpus, the instructed Qwen3-Embedding-8B dense encoder reaches an MRR@10 of 0.849, the best result of any non-decomposed system. Under per-law scoring, this advantage reflects Family Law only
benchmark - arxiv:2610.38335 · cs.LGLearning a Prior-Data Fitted Network for Graph Foundation Models via Structural Causal Model PriorsZiwei Zhang, Tianyin Liao, Chunyu Hu, Ge Lan +6
Graph Foundation Models (GFMs) have recently emerged as a promising paradigm for learning transferable representations across different graphs and domains. However, existing GFMs face notable limitations, including limited pre-training scale or reliance on aligning with textual information in Large Language Models (LLMs). To overcome these limitations, recent studies have drawn inspiration from prior-data fitted networks (PFNs), which learn from large synthetic datasets and support multi-domain prediction backed by causal learning. However, existing graph PFNs encounter severe difficulties in modeling causal relationships within and across node and edge sets, generating diverse graph structures, and designing suitable architectures that can learn from complex synthetic datasets. To address these challenges, we propose **GSCM-PFN** (Graph SCM-based Prior-Data Fitted Network), a graph PFN pre-trained on synthetic data generated from an SCM-based prior that jointly models nodes and graph
benchmark - arxiv:2610.38334 · cs.LGOptimal Regret for Communication-Efficient Decentralized Online LearningSifan Yang, Wenhao Yang, Wei Jiang, Yuanyu Wan +1
In this paper, we investigate decentralized online convex optimization with compressed communication, where $n$ learners over a network jointly minimize a sequence of global loss functions based only on their local losses and compressed messages from their neighbors. The best known regret bounds for this problem are $O(ω^{-1/2}nρ^{-1}\sqrt{T\ln n})$ and $O(ω^{-1}nρ^{-2}\ln n\ln T)$ for convex and strongly convex loss functions, respectively, where $ρ\in(0,1)$ is the spectral gap of the communication matrix and $ω\in(0,1]$ is the compression quality factor. Compared with the corresponding lower bounds of $Ω(ω^{-1/2}nρ^{-1/4}\sqrt{T})$ and $Ω(ω^{-1}nρ^{-1/2}\ln T)$, these upper bounds exhibit polynomial gaps in their dependence on $ρ^{-1}$. To close these gaps, we first develop Ascent-gossip, which combines an online compression strategy with a Chebyshev recursion and attains a faster consensus rate than existing compressed gossip methods. Furthermore, by adopting Ascent-gossip, we propo
online learning - arxiv:2610.38333 · cs.CLFu-TCM: a multimodal large language model that learns traditional Chinese medicine diagnostic reasoningJiahang Xie, Tong Jin, Lingfei Wu, Yaping You +8
Large language models (LLMs) are advancing in multimodal understanding and reasoning, accelerating medical model development. Traditional Chinese medicine (TCM) uses evidence from the four examinations to determine disease location and nature, differentiate syndromes and guide treatment. This diagnostic process constitutes a complete chain of reasoning. We present Fu-TCM, a multimodal TCM LLM trained on data retained after cleaning and filtering an initial collection of approximately 7.12 million examples derived from classical texts, textbooks, medical images, clinical cases and publicly available datasets. The model was further optimized using Bianzheng-Grounded Policy Optimization (BGPO), which explicitly targets syndrome differentiation through rewards for response format, intermediate TCM reasoning fields and final syndrome predictions. Fu-TCM-27B led all six benchmarks spanning TCM reasoning, text and vision, achieving a macro-average accuracy of 81.01% and surpassing the stronge
benchmark - arxiv:2610.38332 · cs.LGMUSE: Multi-Level Memory with Selective Runtime Injection for Scientific Coding AgentsHaoran Jin, Defu Lian, Xiting Wang, Hong Xie
Historical experience can guide scientific coding at multiple levels, but task-level retrieval may miss implementation lessons that become relevant during execution. We introduce MUSE, a memory management framework combining multi-level memory with selective runtime injection for scientific coding agents. MUSE builds global, approach-level, and implementation-level memory from historical trajectories. At initialization, the agent receives global memory and problem-matched approach-level memory. During execution, current interactions are used to retrieve implementation-level memory. An LLM gate selects relevant experience and generates specific review advice, or skips injection. Across four benchmarks spanning numerical problem solving, kernel optimization, and scientific software repair, MUSE outperforms retrieval and memory baselines on all evaluation metrics. Ablations support the multi-level design and gated injection, while transfer experiments show that memory remains useful acros
memoryagentbenchmark - arxiv:2610.38329 · cs.LGA Case Study of a Multi-Agent Kernel-Optimization Loop That Learns a New AI AcceleratorByonggon Chun
Frontier large language models write correct CUDA kernels for most single-operator benchmark tasks, yet their accuracy collapses on accelerators whose architecture and programming language they have rarely seen. We present a case study of an LLM agent loop built for such accelerators. Our example is the FuriosaAI RNGD NPU, programmed in a recent, sparsely documented Rust DSL, where the loop optimized three Gemma-4-12B decode kernels over 12 days on a shared, scarce device. The loop builds its own target knowledge: it writes and checks a summary of the DSL documentation, measures the chip with microbenchmarks, records compiler-derived legality rules, and revisits this knowledge when the toolchain changes, rechecking rules and measurements where it can. A role-based search over per-kernel Design Trees explores whole-kernel designs and accumulates small accepted changes in the best version of each kernel. Because device time is scarce and readings are noisy, cheap checks screen new design
agentllm agentmulti-agentbenchmark - arxiv:2610.38326 · cs.AICan LLMs Faithfully Solve Ontology Engineering Tasks? A Mechanistic Interpretability StudyBohui Zhang, Albert Meroño-Peñuela, Elena Simperl
Ontological knowledge underpins large language model (LLM) performance on knowledge-intensive tasks. However, existing evaluations of LLMs' ontological understanding largely rely on behavioral correctness, leaving the underlying mechanisms unexplained. Existing mechanistic interpretability work has largely focused on instance-level factual knowledge, leaving ontological knowledge under-explored. To bridge this gap, we introduce MIOE, a causal-abstraction benchmark that evaluates interchange interventions for two ontology engineering tasks: entity typing and subsumption prediction. MIOE tests whether learned internal intervention targets can change an ontology-level target answer while preserving class-specific auxiliary information. Across Qwen3-8B and Qwen3-4B, we evaluate SAE-, DBM-, and DAS-based methods, including multi-task and multi-layer variants. Distributed subspace methods give the strongest results and generalize across prompt settings, but the induced changes remain only pa
benchmark - arxiv:2610.38320 · cs.CVSSQA: Sibling-Selective Quadtree Attention for Hierarchical Modeling in Perception TasksYufan Chen, Arnav Bali, Angela Liu, Laura Zheng +1
Perception tasks for navigation in robotics, including aerial platforms such as drones and autonomous driving systems, are inherently structured. Drone-mounted cameras typically capture sky above, terrain below, and obstacles or man-made structures in between, while driving data often contains organized road layouts, lane markings, and surrounding agents. Motivated by these axis-aligned structural priors, we note that such information is typically more structured than in generic image tasks. We hypothesize that processing information in a quadtree-esque manner can not only model features effectively in a hierarchical manner, but also offers an efficient linear-time alternative to vanilla attention mechanisms, which run in quadratic time. In this paper, we propose Sibling-Selective Quadtree Attention (SSQA), which models image tokens hierarchically as a structured, full quadtree. We show analytical complexity analysis that guarantees linear-time feature modeling, in addition to empirica
benchmark - arxiv:2610.38318 · cs.ROWhat Should a Navigator Remember? Spatiotemporal Memory for Zero-Shot Vision-Language NavigationYuhang Peng, Zhikai Wang, Ziyi Wang, Jiaru Zhong +4
Zero-shot vision-and-language navigation (VLN) requires an agent to use past observations without navigation-specific training. We study what information an 8B-scale navigation system should retain and how it should be organized for subsequent decisions. Our agent combines instruction progress, hierarchically compressed observation history anchored to poses, and accumulated geometry in complementary task, temporal, and spatial memories. An 8B observer summarizes observations, and an 8B reasoner uses the memories to select waypoint intents or request additional views, a geometric controller executes the selected targets. A semantic information perspective motivates the distinction between retaining evidence, making it accessible to the reasoner, and acquiring further observations. On R2R-CE val-unseen, the complete system achieves the highest among the evaluated Qwen3-based zero-shot baselines. Removing temporal memory reduces SR and sharply increases average path length from 16.9 to 39
memoryagent - arxiv:2610.38315 · cs.AILLM-PowerTrace: A Dataset of Time-Resolved GPU Power Traces for AI ComputeJiansheng Wang, Jaskeerat Singh, Haochen Xu, Xinyi Li +5
AI compute is becoming large and fast-changing electrical loads. Its impacts on data center infrastructure and power grids depend not only on total energy consumption, but also on how their power draw moves over sub-second to minute timescales, including peaks, ramps, and synchronized fluctuations across accelerators. Existing energy benchmarks for large language models (LLMs) report aggregate quantities such as joules per token and rarely release the underlying time-series measurements, leaving these power dynamics poorly characterized and difficult to model. We introduce LLM-PowerTrace, an open dataset of time-resolved, per-GPU power traces for LLM fine-tuning and inference, sampled every 100\,ms on varying NVIDIA platforms paired with complete workload and system metadata. For fine-tuning, LLM-PowerTrace profiles open model families spanning a wide range of model sizes and fine-tuning methods, together with controlled sweeps over batch size, sequence length, precision, parallelism,
benchmark - arxiv:2610.38308 · cs.LGDoes 3D Pose Improve Action Valuation in Football? An Empirical Study of Pose Representations and Learning ObjectivesRikuhei Umemoto, Kenjiro Ide, Soham Majumder, Ulf Brefeld +1
Football action valuation estimates players' contributions to future scoring. Existing multi-agent deep reinforcement learning (MADRL) models use 2D positions and velocities, omitting posture information relevant to imminent actions. We investigate whether adding 3D pose to a SARSA-based model improves observed-action values relative to attack returns: goals, conceded goals or final expected possession value. Using UEFA Champions League 2022/23 event, tracking and 3D pose data, we compare six representations (no pose, raw joints, torso and head orientation, and spatial-temporal graph convolutional network (ST-GCN) embeddings) across five learning settings and three seeds. The settings vary temporal-difference targets and action-prediction loss, with Monte Carlo return regression as a reference. Across 50 on-ball and off-ball comparisons, pose robustly improved return prediction in only three, all using torso and head orientation, while worsening it in twelve, mostly with ST-GCN. Traini
multi-agent - arxiv:2610.38304 · cs.AIInner-Learning: From Hand-Coded Beliefs to Declarative Knowledge in Trick-Taking GamesStefan Edelkamp
We present a declarative approach to belief-state modeling in trick-taking card games. Starting from a state-of-the-art Skat player, we automatically extract its hand-coded inference rules into a domain-specific language built on the primitive $\neg\mathsf{holds}(P,C)$, denoting that player $P$ cannot hold any card in set $C$. Round-trip code generation verifies that the extracted rules reproduce the original engine's intended behavior, up to three audited bugs in the original that the declarative engine corrects. We then grow the rule base through symmetric derivation and annotation of an expert strategy document---unifying code, symmetry, and expertise in one representation. We extend the knowledge representation with \emph{cardinality constraints}: at-most-$k$ and at-least-$k$ bounds on the cards a player holds from a named group, expressed in the same declarative language and enforced at world-generation time. Confidence parameters connect symbolic rules to probabilistic belief upd
self-play - arxiv:2610.38293 · cs.AIEPSBench: Benchmarking LLM Agents for Earnings ForecastingYanbin Yin, Chuyu Xu, Liqin Ye, Jing Peng +1
Quarterly earnings forecasting requires an agent to connect financial evidence to the income-statement line items that produce a company's reported earnings. We introduce EPSBench, a benchmark for studying this ability across 149 firm-quarters from 86 US companies. Agents read a fixed collection of filings, earnings-call transcripts, and news, then estimate company-specific income-statement drivers with cited evidence. A shared accounting calculator converts these estimates into GAAP diluted earnings per share (EPS), providing a common basis for comparing forecast assumptions. We evaluate six models under a common harness, assessing point forecast error relative to date-matched analyst consensus and checking whether their prediction intervals achieve the stated coverage. Three models show no detectable paired difference in point forecast error from consensus, and three have higher error. Across all six models, nominal 80% intervals contain reported EPS in only 37--70% of runs that retu
agentllm agentbenchmark - arxiv:2610.38292 · cs.AIVendster: An Evaluation Arena to Investigate How Forecasting Tool Quality Affects LLM Agent Behaviors in Buyer-Vendor NegotiationCharin Polpanumas
Time-series forecasters are ranked almost entirely by accuracy metrics, but a forecast's decision value, its effect on the downstream decision, need not track accuracy and is rarely measured. This problem compounds as large language model (LLM) agents increasingly consume forecasters as tools, mediating them through stochastic, opaque policies. We present Vendster, an evaluation arena measuring how a forecaster's bias and prediction-interval width change the outcome of a buyer--vendor discount negotiation between two LLM agents. In the reference instance Vendster-M5, sampling base demand patterns from the M5 dataset, we evaluate two frontier-closed LLM families (Claude Opus 4.6 and GPT-5.6 Sol) and two open ones (DeepSeek-V3.2 and Qwen3-235B-A22B). Model-generated forecasts raise sales target attainment by up to +17 percentage points (p<0.001), yet standard forecast evaluation metrics such as relWQL and MASE do not predict negotiation outcomes at a statistically significant level. To i
agentllm agentarena - arxiv:2610.38291 · cs.ROCompositional Diffusion Planning with Chunk Selection and Reachability MeasureChanyoung Park, Youngjoong Kim, Jaesik Park
Compositional diffusion planners have demonstrated remarkable success on long-horizon diffusion planning by sampling short trajectories, namely chunks, and composing them into a long trajectory. However, averaging overlapping chunks can induce the mode-averaging problem, resulting in discontinuous trajectories. Several works aim to address this problem, whereas a non-negligible portion of plans remains infeasible. In this work, we propose Selective Compositional Diffuser (SCD), handling this problem by measuring reachability. To this end, we present a self-supervised learning method for the reachability classifier. This is trained on offline data and generated chunks jointly to address potential distributional mismatch, while reducing the bias inherited from the generated data. Based on this, we construct a graph whose edge costs represent whether consecutive chunks are reachable, and traverse the graph to find effective yet efficient routes. Lastly, we generate bridging trajectories a
humanoid - arxiv:2610.38289 · cs.CVPulmoWorld: An Asynchronous Multimodal Medical World Model for Longitudinal Pneumonia DynamicsCheng Tang, Xinyu Zhang, Jiawei Chen, Minghui Zhang +13
Pneumonia changes across respiration, inflammation, and organ function, while the measurements that describe it arrive at different times and are often incomplete. We introduce PULMOWORLD, a world model for longitudinal pneumonia dynamics that connects asynchronous state estimation, recursive simulation, and evidence-constrained planning. History-aware filtering separates state propagation from observation-driven correction, so a measurement updates the shared belief only when it is observed. That belief supports recursive forecasts of clinical values and modality representations, and summaries of future prognosis. Cap-MBRL searches imagined continuations inside an authorized action set and replaces a frozen fallback only when a candidate meets the risk budget and a predicted-cost margin. On the pneumonia cohort MASS, recursive forecasts are competitive with task-trained sequence models, and simulated future states retain prognostic information. Trained separately on public MIMIC-IV-CX
world model - arxiv:2610.38273 · cs.LGNAGO: Noise-Aware Generative Operator via Flow MatchingGuanyu Chen, Shengze Xu, Pengwei Liu, Xingyu Ren +4
We introduce NAGO, a noise-aware generative operator for efficient sampling from invariant measures of stochastic partial differential equations (SPDEs) in function space. Our central idea is to treat the source measure as a design variable rather than a fixed Gaussian default. We establish admissibility conditions for well-defined infinite-dimensional flows and show that the finite-band source--target covariance ratio controls the spectral stiffness of the marginal velocity field. This analysis yields an admissibility-constrained spectral-matching rule: choose a source that is sufficiently rough for well-posedness, then minimize covariance mismatch over the resolved Fourier modes. On three representative stochastic-field benchmarks, this source design improves few-step spectral accuracy and several complementary distributional and physical statistics. In a controlled Matern-field ablation, it reduces the accurate RK4 budget from about 30 steps for Gaussian functional flow matching to
benchmark - arxiv:2610.38266 · cs.RORecursive Self-Improvement in Physical AI: Mechanisms, Evidence, and EvaluationXiaofeng Han, Kaize Luo, Zenghuang Fu, Zichong Zhu +5
Embodied systems can update policies, world models, skills, and executable programs from deployment experience, allowing one interaction to benefit later tasks. When a system also revises and reuses its data-acquisition, feedback, or learning procedures, experience changes the process that produces capability, introducing recursive self-improvement. This paper surveys research in physical AI through the connection between persistent capability and its improvement process. We first explain the functional distinction between learner updates and improvers, then compare policy learning, imagined experience, memory and skill accumulation, and executable harnesses. These approaches connect to revisions of curricula, rewards, and research programs in different ways. Contact validity, recovery, retention, and resource costs determine whether the resulting gains accumulate during continued deployment. An L0-L3 rubric distinguishes task-local adaptation, persistent capability gain, inherited pro
embodiedworld modelmemoryself-improvementbenchmark - arxiv:2610.38256 · eess.SYBelief-Driven Search in Connectivity-Constrained Mobile Sensor Networks with Convergence GuaranteesGeorge J. Ziavras, Ziqiao Zhang, Shreyas Sundaram, Christopher G. Brinton
Multi-agent search and coverage is critical to autonomous exploration, search-and-rescue, and defense applications. This paper investigates settings where a mobile sensor network is tasked with coordinating to discover the position of a stationary target within a bounded region. We develop a Bayesian update scheme where agents iteratively combine local uncertainties derived from their sensor measurements through inverse variance weighting to get a better global understanding of where the target may be located given some prior belief. Network agents' trajectories are governed by a primal-dual gradient update rule that simultaneously minimizes the expected sensing uncertainty and enforces network connectivity among the agents. We prove that the belief converges to the target under mild, planner-agnostic conditions, with an explicit exponential rate of false-hypothesis rejection and a bound on expected search time. Simulation results show that we can successfully find the target and colla
multi-agent - arxiv:2610.38255 · cs.LGParaTTT: Parallelizing Test-Time Training Layers with Deep MemoriesNoel Loo, Neehal Tumma, T. Konstantin Rusch, Mathias Lechner +2
Test-time training (TTT) layers replace the linear fast-weight memory of modern recurrent architectures with a small neural network trained online by gradient descent. This nonlinear memory is more expressive than the linear memories of Gated DeltaNet and Mamba-2, but is difficult to parallelize: existing implementations such as TTT-MLP rely on sequential, block-wise weight updates that run several times slower than chunkwise-parallel linear recurrences. Moreover, their block-wise execution only approximates the token-by-token online recurrence, whereas linear recurrences admit an exact chunkwise-parallel evaluation. We introduce ParaTTT (PTTT), an algorithm for parallelizing arbitrarily deep nonlinear memories by iteratively refining estimates of the inner-loop gradient trajectory. Starting from a guess of the hidden-layer gradients, each refinement consists of three chunkwise-parallel passes: a forward pass that evaluates the resulting feature trajectory, an exact gated-delta-rule pa
memory - arxiv:2610.38252 · cs.AIA Tutorial on Generative Models for Music AudioMateusz Modrzejewski, Joanna Kurkowska, Dorien Herremans
Music audio generation spans instrumental clips, full songs with vocals, edits of existing recordings, and systems that play alongside a musician in real time. This monograph explains the methods behind these systems for readers familiar with neural-network training who are new to music generation. It introduces neural audio codecs and autoencoders, autoregressive and masked token models, diffusion and flow matching, sampling acceleration, and preference optimization in a shared notation. Worked examples connect training objectives and sampling procedures to published systems. The application sections relate the musical requirements of instrumental generation, songs, editing, and live performance to model design. Comparison tables record representations, methods, inputs, tasks, and release information documented in papers and official sources. Symbolic composition, speech, and sound effects are included where they connect to music audio. We explain how to interpret automatic metrics an
post-training - arxiv:2610.38249 · cs.CLKeepace.ai: A Technical Report on a Large Language Model for Sports and HealthKeepace. ai Team, Bowen Wu, Heng Wang, Jufeng Mi +11
Large language models hold promise for delivering safe, scientifically grounded, and personalized exercise guidance at scale, yet the sports-and-health domain still lacks curated training corpora and rigorous evaluation benchmarks. Keepace.ai is a domain-specialized LLM built on Qwen3.5-27B through post-training, currently optimized for Chinese-language exercise guidance across three tasks: workout plan generation, sports-and-health knowledge QA, and personal sports data analysis. In internal rubric-based evaluations designed with domain experts, Keepace.ai outperforms its base model and larger general-purpose baselines, with the most pronounced improvements in exercise safety, boundary compliance, structured output completeness, and scientific accuracy. This report summarizes the current design, data pipeline, safety-first evaluation framework, and evaluation results of Keepace.ai.
post-trainingbenchmarkevaluation framework - arxiv:2610.38248 · cs.ROMATCH: Execution-Aware Hierarchical Reinforcement Learning for Vision-and-Language Navigation in Continuous EnvironmentShuhao Ye, Yufei Wei, Demao Wang, Shichao Zhai +2
Vision-and-Language Navigation in Continuous Environments (VLN-CE) requires instruction-consistent planning and reliable execution under collision constraints. Yet many hierarchical approaches optimize the planner while keeping execution fixed. We introduce MATCH, a framework that jointly learns a language-conditioned planner and a goal-conditioned primitive-action executor. An execution-aware topological graph represents start-dependent execution options and integrates executor value estimates with realized outcome memory, enabling the planner to select execution requests and refine their targets through bounded continuous offsets. A transition-aware dense macro reward balances ordered progress along the reference route against execution residual, encouraging executable requests without substituting local success for instruction following. A staged curriculum culminates in synchronized on-policy updates, with the executor learning on planner-generated requests and the planner adapting
benchmark - arxiv:2610.38244 · cs.AIPrecedent or Intent? What Coding Agents Follow When Evolving Existing CodeHaichao Zhu, Geyang Xu, Zhaorui Yang, Jiyuan Wang +1
Existing code reveals past implementation decisions but rarely explains the reasoning behind them. When coding agents implement new features or modify existing functionality, past decisions may conflict with new task instructions. This raises a fundamental question for agentic software evolution: when multiple implementations satisfy the same functional requirements, how do coding agents decide whether to follow existing code or new task instructions? We investigate this question through bidirectional, controlled experiments in purpose-built repositories, followed by validation in real-world open-source repositories. Our approach isolates three competing sources of guidance: (1) agents' default preferences, (2) repository precedents, and (3) explicit task instructions. We conduct 5,625 trials across four coding agents and six implementation pairs, with validation in four open-source repositories spanning three programming languages. We find that agents share the same default implementa
agentic - arxiv:2610.38241 · cs.CVMEV: A Multi-Event Video Dataset for Long-Take GenerationPeiyuan Zhu, Shaoan Xie, Yifan Shen, Wen Tian +3
Current datasets for text-to-video generation are largely limited to short clips depicting a single event or coarse-grained captions. As a result, they provide limited supervision for modeling long-take temporal dependencies, often leading to temporally inconsistent generations when describing multiple events. In this paper, we introduce MEV, a multi-event video dataset designed for long-take video generation. MEV provides high-quality real-world videos where subjects can perform a sequence of distinct, non-overlapping events, featuring precise temporal boundaries and fine-grained procedural annotations. This event-centric discretization reduces confounding factors and enables models to support temporally localized conditioning and build long-take videos from atomic event units. Furthermore, we introduce a comprehensive annotation pipeline that segments videos into temporally grounded events, and generates structured captions at both the global level and the event level. In addition, u
evaluation protocol - arxiv:2610.38239 · cs.AIThe Indian Financial Automated Advice Bias (IFAAB) Dataset : Auditing Socioeconomic and Demographic Disparities in LLM-Generated Financial Advice in IndiaKriti Mahajan
Large language models are increasingly used for financial guidance, yet little is known about whether their advice changes across social identities in Global South contexts. We introduce the Indian Financial Automated Advice Bias (IFAAB) benchmark, a counterfactual audit dataset of 31,734 GPT-4o-generated financial-advice responses for synthetic Indian personas varying by gender, caste, education, employment, residence, age, and district. Holding the underlying request fixed, we evaluate disparities in LLM-implied wealth, income, savings, projected gains, textual similarity, advice topics, and qualitative bias judgments. In the gender-only audit, male-labelled profiles receive 60.1\% higher initial total resources and 46.8\% higher post-advice total resources than comparable female-labelled profiles; in the caste-augmented audit, the post-advice total-resource gap remains 49.4\%. Adding district information reduces the aggregate post-advice total-resource gender gap to 7.9\%, revealing
benchmark - arxiv:2610.38238 · cs.LGVariational Consequence-Driven Offline Reinforcement LearningWen Jiang, Ke Jiang, Xiaoyang Tan
Offline reinforcement learning (RL) typically mitigates distribution shift by imposing divergence constraints on the action distributions of the learned and behavior policies. However, this paradigm cannot ensure the underlying dynamic consistency structure: even minor action deviations can lead to drastically different transitions, inadvertently driving the agent away from the offline dataset. To address this, we explore policy constraints through their induced state-consequence distributions rather than relying on pure action-distribution matching. Nevertheless, directly enforcing such consequence-aware constraints is computationally intractable without access to the true underlying dynamics model. To bypass this issue, we propose Structural Behavior Regularization (StBR), which replaces the intractable objective with a closed-form latent divergence. Structuring this latent space to reflect transition dynamics renders the latent divergence a theoretically bounded surrogate for the tr
agentbenchmark - arxiv:2610.38234 · cs.ROSimRetract: Zero-Shot Sim-to-Real Robotic Assistance for Retractor Take-Over in Open SurgeryXucheng Wang, Zhizhou Yang, Sung Eun Kim, Romain Hardy +6
We demonstrate the first zero-shot sim-to-real transfer for a surgeon-robot collaboration task in open surgery: retractor take-over. A robot arm takes a retractor from the surgeon's hand and preserves retraction at handoff, using a policy trained only on simulated demonstrations, without human demonstrations or real-world fine-tuning. Imitation learning from real demonstrations is limited by collection and labeling costs, and by performance degradation when the environment changes. Our framework, SimRetract, generates simulated demonstrations designed for transfer: the policy observes masked instrument images and grasp geometry, learns to recover from off-nominal states, and acts with margin for execution drift. Deployed zero-shot, it grasps the retractor in 90% of 50 surgeon-defined phantom presentations and preserves retraction at handoff in 78%. Results on unseen retractors suggest potential for generalization to instruments with similar grasp geometry. Across ten policy configurati
sim-to-realgrasp - arxiv:2610.38233 · cs.ROHERON: Real-Time Obstacle-Aware Reachability Maps for Mobile ManipulatorsG. Dupoiron, F. Michaud, J. P. Roberge
Mobile manipulator robots rarely reposition their base during manipulation, because doing so requires real-time knowledge of how well the arm can reach its surroundings among obstacles. Reachability maps (RMs) encode this knowledge, but existing methods are slow, coarse (typically 5 cm), or ignore real obstacles. This article asks whether a neural network can predict an obstacle-aware RM at 2 cm resolution from a live depth camera, on embedded hardware, fast enough for online use. We present HERON, a 3D U-Net trained on RMs computed with the GPU-accelerated inverse kinematics of cuRobo, first on synthetic scenes and then on real depth data, and deployed on a Jetson AGX Orin with TensorRT and INT8 quantization. HERON predicts the RM of a 2.56 m cube around the arm in 115 ms on average, for a measured system rate of 8.57 Hz, close to the 10 Hz update rate of the obstacle map, whereas computing the same map with cuRobo takes about 90 min. Against the cuRobo ground truth, 95% of the voxels
manipulationmanipulator - arxiv:2610.38231 · cs.CV3DCodeEdit: Benchmarking Instruction-Guided 3D Program EditingBinjie Wang, Xinran Zhao, Shoubin Yu, Xiang Li +7
Language models increasingly create 3D content through code, with practical workflows often relying on editing existing programs rather than regenerating them, for speed and precise control. Such edits require understanding what a program builds and foreseeing how a change reshapes it, yet prior benchmarks mostly evaluate generation or simple edits against a reference. In this paper, we introduce 3DCodeEdit, a benchmark of 2,592 instruction-guided editing tasks over 108 Blender programs that construct objects entirely in code. Its edits range from parameter to structural changes, style redesign, and scene composition, and are judged from renders against human-reviewed checklists. Across 17 models, the average success rate is only 37%, and even the strongest models complete fewer than 60% of structural and style edits. As models become more capable, their failures shift to geometric foresight, where the code runs but builds the wrong shape. We therefore study whether multi-turn feedback
agentbenchmark - arxiv:2610.38228 · cs.LGThe Endgame of FSDP Communication: Layout-Aware Collectives on Copy Engines, RDMA, and NVSwitchZejian Wang, Yifan Zhao, Kaihui Ren, Xiaoling Zhou +10
Fully Sharded Data Parallelism (FSDP2) maintains parameters and gradients as per-parameter tensors, whereas its collective path operates on flat, rank-major buffers. Converting between these layouts copies all-gather outputs into parameter storage and packs gradients into full-group fp32 buffers. Layout conversion alone accounts for 6.4% of step time in Llama-3.1 70B training at 8K tokens per GPU. Communication kernels impose a second cost: a default NCCL reduce-scatter can slow a concurrent GEMM that would otherwise complete in a single wave by 69%. We present TIDE, a set of zero-copy collectives built on public NCCL 2.31+ APIs. TIDE preserves FSDP2's per-parameter storage and compute kernels while folding layout placement into collective addressing. TIDE-AG delivers shards directly to parameter storage using NVSwitch multicast and paced recursive-doubling RDMA, with no SMs used for transport or placement. TIDE-RS consumes per-parameter bf16 gradients directly and reduces them into fp
memory - arxiv:2610.38225 · cs.AIFACTS-MAS: A Multi-Agent Architecture for Adaptive Forecasting from Heterogeneous EvidenceMadumita Karthikeyan, Umang Rajnikant Bid, Rong Pan
Real-world forecasting often requires integrating heterogeneous evidence whose predictive value varies across horizons, operating regimes, and domains. We introduce FACTS-MAS, a domain-agnostic multi-agent forecasting framework that decomposes evidence into specialized channels for temporal dynamics, exogenous context, seasonality, structural information, and discrete events. Each agent generates an independently testable forecast through a common interface, while leakage-safe historical validation measures its predictive skill relative to a naive reference. These skill estimates are used to weight and combine agent forecasts, so that uninformative evidence is either removed or downweighted rather than forced to participate. A calibration procedure selects domain-specific constant means based on historical validation folds. We evaluate FACTS-MAS on weekly for-sale housing inventory from Zillow and weekly unemployment insurance claims from FRED, for 15 U.S. metropolitan areas at 4-, 8-,
agentmulti-agent - arxiv:2610.38221 · cs.ROCuraEval: Evaluating Demonstration Curation on Mixed-Skill Robot DataJiuyi Xu, Meida Chen, Yang Ye, Yangming Shi
Robot imitation learning uses demonstrations from operators of varying skill, but selecting training data remains challenging. Existing methods score demonstrations using smoothness, diversity, or estimated influence on policy performance, but differences in selection and evaluation protocols make it difficult to determine whether they outperform simple baselines. We introduce \textbf{CuraEval}, a framework combining mathematical analysis of demonstration scores with controlled comparisons against fixed size-matched random subsets and full-data training on mixed-skill robomimic data using Diffusion Policy and BC-RNN. We show that the evaluated PSD score depends only on episode length and squared position magnitudes, and produces nearly the same ranking as episode length on \textit{square} ($ρ=0.999$). With Diffusion Policy, selecting the shortest 50\% of demonstrations improves \textit{square} success from 70.7\% under random selection to 76.5\%. Compared with random selection, our ada
diffusion policyevaluation protocol - arxiv:2610.38220 · cs.AIRubricon: Automated Rubric Generation and Refinement for LLM Evaluation and TrainingUlyana Tkachenko, Hui Wen Goh, Guram Gogia, Jonas Mueller
Rubrics enable reward assignment on tasks without verifiable answers, but are expensive to manually produce from scratch. We present Rubricon, a pipeline that generates discriminative rubrics without requiring human preference labels, rubric annotation, or generator-model training. Using any LLM, Rubricon samples multiple rubrics, selects the most discriminative candidate, adds criteria to capture overlooked requirements or response flaws, and prunes criteria that are misaligned or shallow. Alongside LLM self-critique and test-time scaling, Rubricon uses observed criterion pass rates on example model responses to guide refinement. Across five professional, consumer, clinical, and general-use LLM benchmarks, Rubricon produces rubrics whose scores of frontier LLM responses exhibit 9% greater alignment with labor-intensive expert-authored rubrics, compared to the best alternative synthetic rubric generation method among RRD, QSRG, OpenRubrics, and agentic generation via harnesses like Cod
agenticpost-trainingbenchmark - arxiv:2610.38215 · cs.AIDisparity Laundering: Consensus Without Fairness in Clinical Multi-Agent SystemsXiaoyang Wang, Tianrui Wang, Christopher C. Yang
Clinical large language models are increasingly fielded as multi-agent teams, on the assumption that deliberation averages out individual error. It does not average out bias. Across a full factorial of five model families and four aggregation topologies on 120 real MIMIC-IV emergency cases under counterfactual demographic swaps, no configuration neutralizes the disparity: all 20 cells commit the direction of their agents' independent answers, giving Black patients more acute triage. Every model has a topology that amplifies its baseline gap, but which one varies by model, so no design is reliably safe or reliably amplifying: the disparity is relocated, never removed. Late-stage fixes fail: fairness instructions and demographic masking leave the gap untouched, and per-agent self-debiasing widens it, because the bias is already encoded in the agents' ordinary clinical numbers. We call this disparity laundering; the audits now applied to clinical AI, per model, per output, per version, ca
multi-agentagent system - arxiv:2610.38214 · cs.LGSkeleton Coding: Program Structure as a Specification for AgentsAditya Golatkar, Alessandro Achille, Renos Zabounidis, Vlad Sobal +8
Coding agents are typically provided end-to-end behavioral specifications that the code they produce should satisfy, such as tests and input--output examples. How many such specifications must a developer provide, and which ones, for the agent to produce the developer's intended program? We argue that this cannot be determined a priori, for it depends on the structure of the implementation, which is left to the agent to decide. An agent first commits to a structure---a decomposition into data types, modules, and functions---which fixes a class of candidate programs; behavioral specifications then select a program within that class. A set of tests that suffices for one structure may leave another badly underdetermined, requiring different edge cases or exponentially more examples. Since the structure is chosen by the agent, developers cannot ensure their specification is complete without reverse-engineering it from thousands of lines of generated code. To resolve this, we introduce Skel
agentbenchmark - arxiv:2610.38210 · cs.AIGAAC: Lossless KV Compression for Bandwidth-Efficient LLM Inference on Personal DevicesYipin Guo, Arun M George, Siddharth Joshi
Weight-only, lossy quantization substantially reduces the storage and memory requirements for serving large language models (LLMs), but the growing key-value (KV) cache remains a bottleneck for long-context inference on personal devices. We present Group-Adaptive Additive Codec (GAAC), a lossless KV compression and serving system that exploits exponent redundancy to reduce storage and accelerate attention without introducing KV approximation. GAAC encodes BF16 exponents using calibrated per-channel base-plus-offset codes over contiguous ranges. GAAC adapts code widths across small groups while preserving KV values bit-for-bit. A fused attention kernel reconstructs KV tiles on the fly, avoiding full-cache decompression. Block-wise compression and fixed-capacity slots support incremental cache updates and paged serving. Across five tested models, GAAC achieves 1.405-1.407$\times$ lossless BF16 KV compression. We also study compression performance on low-precision tensors FP8 E4M3 caches
memorylong-context - arxiv:2610.38209 · cs.AIIs it my preference? Role Grounding Attack against Agent's Memory via Social InteractionHaoran Bu, Yiting Shen, Litian Zhang, Ruihan Hu +1
LLM-based agents now share experiences on social platforms and remember new preferences to enhance their performance. However, existing memory systems contain a vulnerability in role-to-principal attribution that attackers can exploit. We identify Role Grounding Misattribution, in which an attacker-designed interaction causes the victim agent to incorrectly record a preference from the attacker as a preference of the victim's user. Specifically, we are the first to formalize Role Grounding Attack (RGA), an attack model that exploits this vulnerability through nested conversation structures. In the memory formation stage, an experience-sharing post quoting a conversation between the attacker agent and its user is published, which explicitly identifies its publisher. The inner conversation expresses preferences from the user of the attacker agent. When the victim agent remembers preferences from this nested structure, its memory loses the context binding and grounds the attacker's prefer
manipulationmemoryagent - arxiv:2610.38204 · cs.CLTwo Birds with One Stone: Skill-Augmented Generation via Unified Retrieval and UtilizationZihan Luo, Jiahui Kang, Wei Zhou, Qilin Fan +2
Skills, as reusable external knowledge, have demonstrated substantial potential to enhance the ability of Large Language Models (LLMs) to solve complex tasks. However, existing skill-augmented generation methods typically rely on separate models for skill retrieval and skill utilization, resulting in disconnected optimization of these two processes. To address this gap, we propose Skill-Uni, a unified reinforcement learning framework that jointly optimizes skill retrieval and utilization. Specifically, we adopt a generative skill retrieval paradigm that enables a single model to both learn how to retrieve relevant skills and leverage them for task solving. To effectively train this unified framework, we introduce a two-stage training strategy consisting of warm-up instruction tuning and GRPO, allowing the model to progressively improve its skill retrieval and utilization capabilities. Extensive experiments on five skill-related benchmarks demonstrate the effectiveness of Skill-Uni, yie
benchmark - arxiv:2610.38199 · cs.ROEmbodied at Sea: Imitation Learning for Multi-Task Underwater ManipulationGreta Perez-Haiek, Luca Macesanu, Kevin Angers, Mohammed Imam +2
Behavior Cloning has demonstrated promising generalization capabilities for robotic manipulation in terrestrial environments. However, marine visual distortions and dynamic currents create a severe domain shift that remains a key bottleneck in adapting generalist architectures to underwater field manipulation. To address this, 1616 episodes of underwater manipulation data were collected in calm and turbulent waters and were used to train 14 generalist Vision-Language-Action (VLA) models (DynamicVLA) with varying proprioception observation configurations. Across three tasks on 1260 gradable on-hardware rollouts in a pool, 10 models outperformed the baseline, with the highest-performing model achieving a 93.3% overall success rate, including partial successes. Results are also introduced from a case study evaluating the performance of a single-task Action Chunking Transformer (ACT) and a generalist VLA (SmolVLA) policy for two tasks in marine archaeology and conservation: grasping and mo
vision-language-actionvlaembodiedmanipulationaction chunkinggrasp - arxiv:2610.38197 · cs.LGConvex-SKM-Net: Low-Latency Block Halfspace Corrections for Neural Convex Constraint SatisfactionHaoyu Zhu, Xiaochen Zhang, Qingchun Hou
Fast inference and strict constraint satisfaction are essential for learning-based decision-making. Existing methods often rely on constraint-specific projections or feasible-set parameterizations, while iterative repair can be costly. We introduce Convex-SKM-Net, a convex constraint-satisfaction layer that forms linearization halfspaces from residual and (sub)gradient oracles without constraint-specific inequality projectors. It applies fixed-depth dual projected-gradient descent (dual PGD) to minimum-displacement block problems using GPU-friendly matrix operations. We jointly learn the predictor and a positive-definite mobility, adapting the correction geometry to improve objective quality. We prove convergence of the outer iterates at fixed dual depth and linear convergence in distance under Slater's condition and bounded subgradients. All our test outputs on three second-order-cone model-predictive control (SOC-MPC) horizons and six power-system convex relaxations satisfy a $10^{-3
benchmark - arxiv:2610.38196 · cs.LGGearUp: DAG-aware Acceleration for Serverless Inference Serving via Slack-guided Pre-loadingQi Qi, Yifan Sui, Hanfei Yu, Hong Zhang +1
Serverless platforms increasingly host machine learning (ML) inference workflows, modeled as directed acyclic graphs (DAGs) of interdependent functions. Beyond container startup, each ML node must load libraries, model weights, and CUDA kernels before it can run. These costs accumulate along dependent DAG paths and place initialization on the end-to-end critical path. Existing systems optimize one half of this problem. DAG-agnostic pre-loading prepares nodes the request may never reach, wasting host and GPU memory. DAG-aware serving routes and scales whole functions, leaving the loading stages inside a node untouched. This paper presents GearUp, a serverless DAG inference serving system that removes ML-node loading from the end-to-end critical path at low resource cost. GearUp profiles the loading stages of each node and schedules them inside the slack time the DAG exposes, hiding initialization behind upstream execution. It reuses initialized instances within and across requests to av
memory - arxiv:2610.38194 · cs.AIPrototype-Guided Dynamic Domain Modulation and a Large-Scale Multi-SKU Benchmark for Shampoo Bottle Surface Defect DetectionYuehang Li, Xiao Wang, Qiang Chen, Yangyi Fan +3
Industrial shampoo-bottle surface defect detection must operate across many stock keeping units (SKUs), where differences in bottle geometry, material, label layout, reflectance, and imaging conditions can substantially alter the appearance of the same defect. Existing industrial defect datasets rarely provide both instance-level defect annotations and product-variant identities, making it difficult to study SKU-aware detection under realistic manufacturing conditions. We introduce Bottle-SKU, a large-scale multi-SKU benchmark for shampoo bottle surface defect detection. Bottle-SKU contains 13,119 images, 20,419 instance-level bounding boxes, seven defect categories, and 48 bottle SKUs, enabling evaluation of both shared defect recognition and SKU-specific appearance variation. To address the challenge of SKU-specific appearance variation, we propose Prototype-Guided Dynamic Domain Modulation (ProDMM). ProDMM constructs a visual prototype for each SKU from training images and predicts
benchmark - arxiv:2610.38180 · eess.SYMachine learning-guided real-time AC optimal transmission switchingWeimin Huang, Weiwei Yang, Bistra Dilkina, Konstantina Mellou +4
The operation of power grids has substantial economic consequences, affecting generation costs and, ultimately, electricity prices. The Optimal Transmission Switching (OTS) problem seeks to reduce these costs by selectively switching off transmission lines. However, OTS gives rise to a challenging mixed-integer nonlinear optimization problem that is difficult to solve at operationally relevant timescales, particularly as rapidly changing loads create a need for real-time switching decisions. To address this challenge, we develop a machine learning-guided method that identifies high quality line switching decisions in real time. Specifically, we construct a grid-specific switching-value model that rapidly estimates the generation-cost reduction associated with opening transmission lines by fine-tuning a pretrained model for AC optimal power flow (AC-OPF). The switching-value model provides signals to guide promising switching decisions, and topologies proposed by the model are evaluated
benchmark - arxiv:2610.38175 · cs.AIDecompGym: Benchmarking LLM Agents on Designing Decomposition AlgorithmsSikai Cheng, Tinghan Ye, Arnaud Deza, Xiaonan Qi +3
Benders decomposition and column generation solve large optimization problems by passing cuts, columns, and dual values between a master problem and subproblems, and are hard to implement correctly. Large language models now automate parts of algorithm design, which points toward agents that build decomposition solvers for new problems. As a first step toward such agents, we study how well current coding agents design and implement these two algorithms, and how they fail. We introduce DecompGym, 81 instances from nine problem families (five Benders, four column generation), in which an agent writes a solver for a prescribed method. Each solver must meet three criteria: execution within a 60-second time limit; numerical feasibility together with a method audit, in which an LLM judge reads the solver's code; and an objective within 1% of the stored optimal objective. We evaluate 12 GPT, Claude, and Gemini models running in coding-agent tools. Over up to ten attempts, we compare independe
agentllm agentbenchmark - arxiv:2610.38174 · cs.AIScaling Lean Theorem Proving for Real-World Code VerificationDebargha Ganguly, Divya Raghunathan, Dejan Jovanović, Vikash Singh +5
We present EasyProver, a Lean proving system built from recursive agents with isolated contexts, asynchronous search, parallel tactic probing, and an elastic elaboration service. Using an off-the-shelf frontier model with no additional training, it proves 669/672 PutnamBench problems and all six IMO 2026 formalizations and resolves 597/600 NTP4VC verification conditions and all 189 VERINA tasks by proof or refutation. Because these benchmarks are close to saturation, we also introduce Aegis, which converts competitive programming problems into much harder verification tasks over real implementations. The pipeline writes safe Rust solutions, translates them into Lean with Aeneas, generates specifications, and labels each frozen statement with a kernel-checked proof or refutation. It produces 2,045 differentially validated Rust-Lean program pairs and 11,533 candidate specifications. Unlike hand-written benchmarks, these tasks carry implementation-level obligations from fixed-width arithm
benchmark - arxiv:2610.38162 · cs.CVNeighbourhood shortcuts in building damage assessment from very-high-resolution satellite imagery: diagnosis and a building-evidence plug-inJianChong Guo, YuTing Wan, Kefan Li, AiLong Ma +1
Building damage assessment (BDA) from very-high-resolution imagery supports disaster response, and deep models now achieve high overall accuracy on public benchmarks. Because damage is spatially clustered, most buildings share the damage level of their neighbours, so overall scores are dominated by such buildings, whereas damaged buildings among intact ones are rare and easily overlooked. A high overall score therefore cannot tell whether a model judges the building itself or relies on its neighbourhood as a shortcut, and this study tests which of the two current models do. We stratify buildings by their agreement with their neighbours, move buildings into new neighbourhoods with their pixels unchanged, and diagnose nine models covering the main paradigms of BDA. All models miss a third to half of the isolated damaged buildings, and in non-flood events they call 73--87% of intact buildings damaged once these are moved into a damaged neighbourhood. We propose a building-evidence plug-in
benchmark - arxiv:2610.38160 · cs.CLATI-OPD: Adaptive Teacher Intervention for Stable Long-Sequence On-Policy DistillationJiannan Guan, Youxin Jiang, Yutai Hou, Qiguang Chen +8
On-policy distillation (OPD) trains a student on its own responses using token-level feedback from a teacher. As responses grow longer, the student's prefix can depart from the teacher's own reasoning, weakening its feedback on later tokens. Updates based on this feedback can reinforce the same deviations in later responses and destabilize training. Prior approaches such as length control and loss reweighting limit learning from problematic responses but leave their prefixes unchanged. Fixed cutoffs can also discard later reasoning. We propose \textbf{Adaptive Teacher Intervention for On-Policy Distillation (ATI-OPD)} to improve subsequent teacher guidance by revising the student's reasoning prefix. ATI-OPD replaces the earliest high-disagreement segment with a short teacher segment, then lets the student resume its own reasoning from the revised text. To handle renewed disagreement, it excludes persistently divergent suffixes from the distillation loss while retaining useful student c
benchmark - arxiv:2610.38158 · cs.CVREEL: Same Frames, Different Answers and Different CostsXin Liang, Jiangpeng He
Video understanding keeps moving toward longer recordings, and the longer a recording is, the more questions can be asked about it. But processing it again for every question does not scale, so many systems watch the video once, build a memory, and answer later questions from it. Yet what to remember is decided before anyone knows what will be asked, and whatever is left out cannot be recovered later. Moreover, whether the savings pay off is judged by token counts, which fall when the input gets shorter but not when the model reuses what it has already computed. We present REEL, an evaluation framework that compares three saved representations built from exactly the same observed frames, so that a change in answers can come only from what the system remembered. Text writes a short record of what happened, Visual keeps the encoder outputs, and Prefix State also keeps the language-model prefix. Each is compared with full reprocessing (Frames) by paired correctness changes, cumulative GPU
benchmarkevaluation framework - arxiv:2610.38155 · cs.LGFloor It, Then Fill It: Simulating Creator Ecosystems for Fair Video-Content Recommendation Systems at Billion User ScaleSohini Roychowdhury, Long Sha, Sunand Agarwal, Payal Kamboj +4
Engagement-optimized content recommendation systems exhibit a well-documented rich-get-richer dynamic that structurally disadvantages emerging creators on large-scale platforms. Evaluating debiasing interventions at scale demands faithful simulation of creator-viewer ecosystems, yet the fidelity, diversity, and validity of such simulators remain underexplored. We present a grounded user-simulation framework that models a heterogeneous population of ${\sim}42$K creators and their coupled viewer interactions over 12~weeks, and use it to evaluate equitable view-allocation policies formulated as a budgeted multi-armed bandit (MAB). The simulator captures three behavioral dimensions: (i)~\textbf{creator diversity}, modeled via a clustering approach grounded in real platform creator-level distributions; (ii)~\textbf{behavioral fidelity}, captured through coupled creator-content state machines that represent growth trajectories, engagement accumulation, and graduation dynamics; and (iii)~\tex
sim2real - arxiv:2610.38154 · cs.LGMASS: Multi-mask Selective Self-Distillation for Leakage-Resilient ReasoningXiuqi Liu, Oveys Delafrooz Noroozi, Upamanyu Madhow
On-policy self-distillation enables the same model to act as student and teacher, with dense token-level supervision obtained by the teacher being given access to privileged information. However, conditioning the teacher on a single reference can concentrate its predictions on one reasoning trajectory and limit transfer to the student. We propose Multi-Mask Selective Self-Distillation (MASS), which approximates the unavailable marginal over diverse reference realizations using multiple independently masked views of one verified solution. MASS aggregates these complementary teacher predictions, selectively transfers reliable token-level corrections, and adaptively focuses masking on influential reference spans. Across mathematical reasoning benchmarks and model families, MASS consistently outperforms strong self-distillation baselines.
benchmark - arxiv:2610.38151 · cs.LGEstimating Implicit Expected Log Density RatiosAvi Amalanshu, Chandler Squires, Yizhou Lu, Pradeep Ravikumar
When exact probabilities cannot be computed, \emph{density ratio estimation (DRE)} is a key subroutine in the wide range of applications where density ratios appear, from experimental design to generative modelling and reinforcement learning. In this paper, we introduce the concept of an \emph{expected log density ratio (ELDR)}: log density ratios that are integrated under some third, possibly distinct measure. We show many problems with implicit densities where DRE is naïvely used are actually ELDR estimation tasks, and prove fundamental limitations on finite-sample ELDR estimation. Based on these results, we argue that the typical DRE-based approaches are unsuitable for ELDR estimation, and provide some foundational motivation towards bespoke ELDR estimators. To demonstrate, we propose a simple modification to some DRE methods that allows samples convergent to the integral distribution to appear in training data. We show through benchmarks that this can improve ELDR estimation error.
benchmark - arxiv:2610.38147 · cs.AISkillCT: A Multidimensional Benchmark and Execution-Free Baseline for Agent Skill AssessmentMengyu Sun, Zunlong Zhou, Ziyuan Yang, Yi Zhang
Agent skills package instructions, scripts, and resources for specialized tasks and are increasingly distributed through public Web repositories. However, existing skill benchmarks primarily focus on downstream task execution, insufficiently assessing specification quality, behavioral robustness, semantic alignment, and responsibility boundaries. Their reliance on task-specific tools, environments, and repeated executions limits scalability and pre-deployment assessment. To address these challenges, we introduce SkillCT, a benchmark that broadens skill assessment from downstream utility to a multidimensional characterization of skill quality. SkillCT evaluates skills across nine complementary dimensions spanning three aspects: Static Assessment, which captures intrinsic properties of the skill specification; Dynamic Assessment, which measures whether the induced behavior remains stable across backbone models and request variations; and Human-Assisted Assessment, which examines behavior
agentbenchmark - arxiv:2610.38145 · cs.MAEl Agente Quntur: Agentic Active-Space Selection for CASSCFJuan B. Pérez-Sánchez, Viviana Roman-Ventura, Olalla Nieto-Faza, Aiwei Yin +2
Multireference calculations often require substantial human expertise and time to choose and refine an active space for the electronic states or chemical process of interest. This involves selecting which orbitals to include, verifying that the calculation describes the intended excited state, and deciding whether the results have converged. Here, we ask whether scientific agentic systems (harnesses) can make these decisions reliably. We use the El Agente Quntur multi-agent harness, with GPT-5.6 Sol and GPT-5.6 Terra as the underlying LLM components, to explore 32 diverse singlet excitations at fixed molecular geometries. The agents planned and executed calculations, choosing and refining active spaces under prompts that specified equal-weight two-singlet state-averaged CASSCF followed by strongly contracted NEVPT2 with def2-TZVPD. We evaluate excitation energies against published CASPT2 reference values at matching geometries and compare performance with six deterministic Active Space
multi-agentagentic - arxiv:2610.38141 · cs.ROSKiD-SLAM: Resource-Aware Distributed Multi-Robot LiDAR SLAM for Planetary Missions with Field Deployments at Analog SitesHogyun Kim, Juwon Kim, Dongjin Cho, Jiwon Choi +6
Future planetary exploration missions require teams of robots to collaboratively map unknown environments under severe communication limits, constrained onboard computation, and no central infrastructure. In this field report, we present SKiD-SLAM, a distributed multi-robot LiDAR simultaneous localization and mapping (SLAM) system for such environments, and report its deployment across two mission concepts: cave exploration and scientific surface exploration. We survey the communication and computation budgets of actual planetary missions and select modules that fit these budgets for inter-robot place recognition, relative pose estimation, and distributed trajectory fusion. Before field deployment, we verify the assembled system on public datasets resembling planetary environments and in a simulator built on NASA JPL's Open Source Rover. We then validate SKiD-SLAM through field deployments in South Korea: a team of wheeled rovers explores the Majang lava-tube cave on Jeju Island, and q
quadruped - arxiv:2610.38127 · cs.LGA Separation Principle for Adversarial Attacks and Defense of Graph Neural NetworksFarid Mafi, Saber Jafarpour, Mohammad Pirani
Adversarial attacks on graph neural networks (GNNs) have been studied as a model-centric optimization problem driven by gradients, learned policies, or meta-learning, with the graph treated as a passive substrate. This view overlooks a defining feature of GNNs: their predictions are produced by structured message passing along the graph, so their adversarial vulnerabilities are encoded as much in the geometry of the graph as in the learned weights. We develop a structure-driven framework for feature attacks in which the attack is derived from the spectral properties of the normalized adjacency operator. Its central result is a separation principle: under a natural relaxation of the attack objective, the optimal feature perturbation is rank one and factorizes into a graph-dependent component (degree-proportional spatial allocation, depending only on the graph) and a weight-dependent component (top singular vector of the composed weights, depending only on the weights). Node degree thus
benchmark - arxiv:2610.38121 · cs.LGWhat Does LOL Measure? Hidden Label Shifts in Low-Light Enhancement BenchmarksAdityan Thirumani Srinivasan
LOL-v1 and LOLv2-Real are the standard benchmarks for supervised low-light image enhancement, and several leading methods ship hand-set test-time adjustments with their evaluation scripts. We audit both benchmarks with 18 public checkpoints of eight methods under one metric implementation that reproduces every published number to within 0.09 dB, and find that their test labels differ from their training labels. (i) The 15 LOL-v1 test labels are rendered differently: the same captures appear in LOLv2 with labels of identical luminance but 1.48x lower saturation, and LOL-v1's own training labels match the LOLv2 rendering; we confirm this on the official release. (ii) The LOLv2-Real test set is the contiguous block of images 690-789, whose labels are 14% darker than the training labels (Cliff's delta = -0.45, p < 1e-12) with no comparable shift in the input statistics we examined; methods trained without any labels score up to 5.5 dB higher on test than on train. (iii) All 15 LOL-v1 test
benchmark - arxiv:2610.38120 · cs.LGPost-Training Large Language Models with LLMONBasel Shbita, Michael Hind, Chad DeLuca, Nathan Fulton +7
Large Language Models (LLMs) treat all context tokens uniformly, lacking explicit separation between executable instructions and passive data. This leads to prompt injection vulnerabilities, ambiguous instruction binding, and unreliable execution control. We present a post-training methodology using LLMON (LLM Object Notation), a framework that enables separating instructions from data and requires explicit identifier-based execution. We construct a large-scale structured training corpus of 4.6B tokens spanning general, RAG (Retrieval-Augmented Generation), and tool-calling data, augmented with distractor-infused variants in which competing instructions and irrelevant data segments are present in context but not referenced. Post-training two Granite-4.1 base models (3B and 8B) through supervised fine-tuning followed by reinforcement learning, we show that LLMON-structured supervision reduces attack success rates on indirect prompt-injection benchmarks from 53-100% to 0.02-5.6% while pr
retrieval-augmentedragpost-trainingbenchmark - arxiv:2610.38109 · cs.LGWho Called? V33DA: A Physically Verified Multimodal Benchmark for Vocal Attribution in Zebra Finch GroupsMaris Basha, Yuhang Wang, Xiaoran Chen, Longbiao Cheng +4
Deciding which member of an animal group produced a vocalization and where that vocalization was produced are central to studying social communication but are rarely evaluated with physically verified ground truth. Existing resources either localize sounds without identifying the caller, or recognize individuals from isolated vocalizations without reliance on the 3D candidate geometry. We close this gap with V33DA, a dataset-centered multimodal benchmark for spatial vocal attribution in social zebra finches: it comprises 33,625 verified events coupling 5-channel audio with three camera views, calibrated 3D pose for every visible candidate, and per-bird radio telemetry across 10 individuals and three experiments. Caller identity is verified from an on-body accelerometer-derived vibration signal; this channel is withheld from benchmark models and used only by an on-body reference ceiling. Given a detected vocalization and the visible candidate birds, the task is to select the caller. We
benchmark - arxiv:2610.38103 · cs.CVThe Prestige: Benchmarking Cognitive Visual Reasoning using Magic TricksShuo Wen, Beixi Du, Edwin Meriaux, Junming Shi +4
We introduce The Prestige, a video understanding benchmark that probes cognitive visual reasoning through professional magic performances, where deliberate misdirection challenges models to track events and explain unexpected outcomes. The benchmark comprises open-ended questions ranging from temporal and causal understanding to robustness against adversarial questions. Evaluation of leading open-source and proprietary VLMs reveals a substantial gap between even the strongest model and human performance. We also find systematic discrepancies between human and automated grading and identify recurring model failure modes. Ablations further show that models can select correct multiple-choice answers even when they fail to generate correct explanations independently, motivating further investigations into question formats. Together, these findings highlight limitations in both VLM reasoning and its evaluation, underscoring the need for more reliable video understanding models and more rigo
benchmark - arxiv:2610.38101 · cs.AIAI-Driven Digital Twin Framework for Warehouse Logistics Automation OptimizationMd Mahinur Alam, Kanita Jerin Tanha, Md Tanjin Alam, Shadman Sakib +1
In recent years, warehouse logistics optimization has increasingly focused on intelligent, data-driven systems to enhance operational efficiency. Conventional static product placement strategies are being progressively replaced by dynamic slotting approaches that can adapt to real-time variations in warehouse conditions. In this work, we present a novel framework that employs a reinforcement learning agent integrated within a FlexSim-based Digital Twin environment for real-time product placement optimization. The proposed agent continuously learns from streaming operational data, enabling it to reduce picking time while improving overall system throughput. The simulation environment models a four-zone warehouse layout with forklift-based material handling and leverages the Asset Administration Shell framework to ensure seamless interaction between the learning agent, physical asset, and the digital twin. Experimental evaluation demonstrates substantial performance gains, with the propo
agent - arxiv:2610.38100 · cs.LGEvaluating Causal Reasoning in LLMs: A Case Study in the Challenges of Procedural GenerationJustin Clarke, Isha Karn, Patrick Murray, KJ Hill +1
Large language models (LLMs) appear to represent and reason about a large range of phenomena. In particular, recent studies have examined whether LLMs can accurately reason about causation. One of the most commonly used benchmarks for evaluating formal causal reasoning in LLMs is CLadder, a framework for procedural generation of benchmark data sets that was released in 2023. Experiments with CLadder appear to indicate that LLMs can perform formal causal reasoning in a large percentage of generated examples. In this paper, we examine what can be reliably inferred when LLMs perform well on test sets produced by CLadder. We identify several examples of superficial and highly predictive patterns that the CLadder generator embeds in data sets it generates. We find that even simple models such as random forests perform well on CLadder-generated data sets after training on a sample of generated data. We also find that the performance of many commercial LLMs on CLadder test sets is extremely b
benchmark - arxiv:2610.38096 · cs.AICalculation and rejection diverge in language models tested at known mechanical limitsShaoliang Yang, Jun Wang
When a language model returns engineering quantities with the status "solved", software that trusts it passes them on, even when the requested design is physically impossible. Such failures could reflect poor calculation, how the answer is requested, or a failure to judge feasibility. We set beam loads and disk speeds at known percentages above or below plastic collapse or burst, measured each model's limit calculation separately and tested 12 models in 4,860 calls, including follow-up experiments on new designs. We find that a model can recognize an impossible design when asked only whether it is feasible, yet accept the same design, reporting "solved", when asked for a full answer. Three models accepted beams 10 to 30% beyond collapse despite precise limit calculations in separate questions, and many accepted replies stated the violation. Asked only about feasibility, Opus 5 and Gemini 3.8 Flash rejected every impossible new design they were given, including those accepted in full an
benchmark - arxiv:2610.38095 · cs.LGNemotron 3.5 Super and Lightning: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Models for Agentic AINVIDIA, :, Aastha Jhunjhunwala, Aayushi Ashokkumar Srivastava +555
We introduce Nemotron 3.5 Super and Nemotron 3.5 Lightning, open and efficient Mixture-of-Experts Hybrid Mamba-Transformer models designed for reasoning, long-context understanding, and agentic workloads. Super is a 120B-parameter, 12B-active multimodal model, while Lightning is a smaller 30B-parameter, 3B-active text model; both designed for efficient deployment. Both models combine large-scale distillation, long-context training, and native speculative decoding, while Super further incorporates Indic continued pretraining and agentic software-engineering mid-training. We develop a comprehensive post-training pipeline spanning supervised fine-tuning and reinforcement learning (RLVR and RLHF), complemented in Super by specialized teacher training and multi- teacher on-policy distillation (MOPD). This transforms Super from a text-only base model into a multimodal agent capable of reasoning over text, images, video, and interactive tool environments. We further explore test-time scaling
long-contextagentagenticpost-trainingbenchmark - arxiv:2610.38090 · cs.AIPredicting What Answer Engines Cite: Open Decision Models, Rerankers and a Distilled 4B Student on Ten Engines under Controlled RetrievalElisha Bajemon, Andre-Louis Rochet
Generative engine optimization needs a scorer that ranks candidate pages by the share of an answer they will receive as citations. We measure pretrained rerankers, a hosted typed decision model (Jev), eight open decision models and distilled students on GEO-Bench queries answered by ten engines that receive the same candidate sources in their prompt. Where several engines answer the same query, they agree with one another at a within-group Spearman correlation of 0.704. Jev and two open 27B decision models, used without training, are the strongest scorers at 0.575 to 0.578. A Qwen3-Reranker-4B fine-tuned with low-rank adapters on teacher labels for other GEO-Bench queries, released as GEO Citation Ranker 4B, reaches 0.542 to 0.549 across three seeds, and its differences with the 27B models and with Jev are not distinguishable from zero on 229 queries. Most of the remaining distance between the student and Jev disappears when Jev scores one passage at a time, as the student does. Studen
benchmark - arxiv:2610.38086 · cs.CVFrom Next Day Wildfire to Next Days Wildfire: A New Multi-Continent Dataset of Active-Fire Spread for Sequence-to-Sequence ForecastingJose Guillen, Renzo Zanca, Valentin Barriere
Benchmarks such as WildfireSpreadTS (WSTS) have enabled data-driven forecasting of wildfire spread from multimodal satellite time series, but they suffer from limitations: size restricts them to large fires that burn for weeks, and they are from similar environments. We introduce AFSeq, a multimodal dataset of 1,194 recent fire events (2023 to 2026) from 8 different countries across three regions, dominated by the short, small fires that existing benchmarks underrepresent. Our dataset is the largest and most diverse to date for time-series-based wildfire spread prediction. We also introduce UTAE-AR, an extension of a classical U-Net with Temporal Auto Encoder, enhanced with recurrent modules before the decoder. This allows the model to not only predict the next day but the whole sequence in an autoregressive way. Moreover, our model can feed its own predicted fire mask back one day at a time, producing the day-by-day forecasts that single-step models cannot. Our results show that UTAE-
benchmark - arxiv:2610.38065 · cs.CVVision Tomography: Towards Portable 3D Bidirectional Face-Skull Inference from Visual DataZidu Wang, Meng Xu, Miao Xu, Hengyuan Ma +4
Traditional computed tomography (CT) can capture paired 3D face and skull geometry, but it remains costly, fixed in location, and exposes patients to radiation. For applications such as remote diagnostics, digital healthcare, forensic identification, missing-person investigation, and archaeological analysis, there is an urgent need for a portable, safe, and accessible solution to capture 3D skull geometry. In this work, we introduce Vision Tomography (VT), a concept and framework capable of reconstructing paired 3D face-skull geometry from various visual modalities, and also enabling joint manipulation of 3D facial and skull geometry. To achieve this, we first construct a dataset of over 200 samples, where each case includes a CT-based skull, a CT-based face, and a high-fidelity textured face scan. Secondly, we propose a novel dense ray matching registration method that ensures topological consistency across face, skull, and their tissue correspondences. Building on this registration,
manipulation - arxiv:2610.38053 · cs.LGCooperative EV Charging Assignment Using Multi-Agent Reinforcement Learning: A Comparison of QMIX and Independent Q-Learning Under Urban Traffic UncertaintyNour-Eddine Moumni, Rachid Alaoui, Driss Kiouach
The allocation of electric vehicles to public charging stations is increasingly constrained by the fact that several heterogeneous facilities often serve the same stream of requests while observing only partial local information. An accept/reject decision that is defensible for one station may nevertheless reduce the efficiency of the charging district as a whole. We formulate this operational problem as a Decentralized Partially Observable Markov Decision Process (Dec-POMDP) and examine whether Centralized Training with Decentralized Execution (CTDE), instantiated through QMIX, improves assignment performance relative to a cooperative Independent Q-Learning baseline (IQL). The evaluation uses a six-station urban network simulated with SUMO, five independent random seeds, 200 training episodes, and four traffic flux scenarios ranging from 100 to 250 vehicles/hour. Under this protocol, QMIX reaches a mean reward per evaluated decision of +0.776, versus +0.578 for IQL (delta = +0.198, se
multi-agent - arxiv:2610.38045 · cs.CVClassifying Urban Locations by Cyclist Facility Usage and Directional Alignment: A Validated Camera-Based Framework from Multi-Site ObservationsAo He, Mohit Gupta, Deborah Salon, Ricardo Eiris +1
Objectives: Agencies invest in bicycle facilities but lack tools to assess whether cyclists use them as designed. We examine whether usage patterns follow facility configurations and whether higher bike lane usage is associated with less wrong-way riding. Methods: Motion-triggered cameras covered 19 deployments at 17 locations in Tempe, Arizona, in 2023. Manual review identified 1,963 cycling events; excluding 20 crossings left 1,943 labeled events for analysis. An automated pipeline on the same captures counted 92.9 percent of the full manual total (per-deployment r = 0.99), and every automated wrong-way call was visually verified. Deployments were classified using thresholds on facility usage shares. Findings: Deployments were bike-lane dominant (9), sidewalk dominant (6), or mixed (4). Cyclists in bike lanes rode against the flow in 4.6 percent of events; on sidewalks, riders opposed the near-side traffic flow in 49.6 percent. Across deployments, bike lane share and wrong-way rate w
benchmark - arxiv:2610.38044 · cs.AIPromptSense: Sentiment-Aware Mental Health Chatbot with Dynamic PromptingSugi Lee, Jianlong Zhou, Jing sUN
Current Large Language Models (LLMs)-based mental health chatbots often lack sufficient empathy, contextual awareness, and safety mechanisms, potentially leading to ineffective or harmful interactions. This paper proposes a novel framework that integrates sentiment and depression analysis with a rule-based filtering mechanism prior to model inference. By dynamically constructing prompts based on emotional and risk-related insights, the framework aims to significantly improve the contextual relevance, empathy, and risk-aware of model-generated responses. We evaluated four model strategies and two prompt engineered variants across qualitative metrics such as empathy, helpfulness, clarity, relevance, safety, as well as toxicity assessments using the Perspective API. Experimental results indicate that the prompt engineering approaches outperforms DSFT and RLAIF by 1.8\% and 1.1\% respectively.
rlaif - arxiv:2610.38042 · cs.CVCan Event Detectors Leave the Broadcast Camera? Evidence from TennisDavid Soeiro-Vuong, Andre-Louis Rochet
Ball bounces and racket hits are the events from which tennis statistics are derived, and on monocular video the two are easy to confuse. We release a benchmark of 345 bounce and hit annotations at frame level on eight clips that mix broadcast, fixed elevated, phone and court-level viewpoints, with a frozen development and test split, a confusion-aware matcher, clip-level bootstrap intervals and a blind second annotation of three clips. On it we score supervised precise spotters with published broadcast weights, two zero-shot vision-language models, and RallyArb, an arbitration layer over a ball track and player poses that fits no parameter on event labels. The supervised spotters score near-perfectly on broadcast footage and almost nothing on phone and court-level footage, where the pooled bounce F1 of E2E-Spot is 0.130. RallyArb keeps a bounce F1 above 0.8 on the three phone and court-level clips, two of which were used to tune it, and finds 8 of 11 bounces on the held-out one, but i
benchmarkevaluator - arxiv:2610.38041 · cs.CLCARF: Context-Aware Reasoning Supervision for Scientific Evidence ScreeningNishtha Agarwal, Adam Visokay, Astha Wahi, Kwabena Owusu Aninkora +3
Systematic reviews synthesize biomedical evidence to inform clinical and public-health decision-making, but require domain experts to manually screen thousands of studies against detailed eligibility criteria. We investigate whether language models can assist with this bottleneck by learning from expert feedback on title and abstract screening decisions and model-generated reasoning. We introduce the Context-Aware Reasoning Framework (CARF), a human-in-the-loop framework that uses expert feedback on model-generated reasoning to adapt language models to context-dependent screening tasks. We evaluate CARF on systematic reviews of verbal autopsy and cause-of-death ascertainment methods used in global public health mortality surveillance. CARF shows competitive screening performance compared to evaluated models while retaining high sensitivity, and is more inclined to retain ambiguous, context-dependent studies for human review. Together, these findings suggest that expert-corrected reason
human-in-the-loop - arxiv:2610.38034 · cs.ROLearning from Relative Action Credit for Escaping Local MinimaGiovanni Minelli, Giulio Turrisi, Lorenzo Amatucci, Victor Barasuol +1
Reinforcement learning holds the promise of robots that learn manipulation skills from interaction alone, without pretrained priors or demonstration data. In practice, on-policy methods such as PPO routinely stall on long-horizon, contact-rich tasks: dense reward shaping carries the policy through stages, yet stitching them together and discovering sparsely rewarded events rarely succeeds. We trace this failure to the missing action discrimination in state-value credit assignment: a state-value critic judges progress only by episode outcomes, so when the policy converges to a partial solution all outcomes coincide, and improvement stops because no action can prove better than another. We therefore extend PPO with an action conditioned value critic which, can compare the executed action with alternatives sampled at the same state and aid the learning signal. Combined with Generalized Advantage Estimation and a bound on exploration noise, we show how the extended algorithm escapes the pl
manipulation - arxiv:2610.38027 · cs.ROMETATRON: A Promptable Whole-Body Execution Layer for HumanoidsChen Tessler, Yifeng Jiang, Alec Otto, Yi Shi +3
Humanoid controllers often require task-specific training despite sharing the need for balance and whole-body coordination. We present Metatron, a promptable whole-body execution layer combining a generative base policy with a programmable skill layer. The base policy receives partial goals, collectively called a prompt, that can combine timed body-part targets, velocity commands, and style cues. It produces diverse whole-body motions consistent with these goals. Training combines motion capture, synthetic recovery trajectories, and retargeted scene-interaction motions. After training, the policy is frozen. Stateful programs called skills express task intent as prompts at each control step, drawing rough geometric sketches that the policy turns into coordinated movement. Skills are ordinary code, refined by coding agents that analyze motion data and fit compact rules without task-specific policy training. On the physical Unitree G1, one frozen checkpoint supports walking, recovery, tel
humanoid - arxiv:2610.38024 · cs.AIAssessing Language Models for Scientific Applications: Inertial Confinement Fusion as a Use CaseP. B. Radha, D. D. Meyerhofer, M. Bhattarai, B. Cheng +10
Advancing language models for specialized scientific applications requires evaluations of scientific fluency, physical reasoning, and quantitative analyses. Using inertial confinement fusion (ICF) as a test case, we have developed an expert-curated benchmark of 250 questions spanning qualitative reasoning, numerical calculation, symbolic derivation, and equation formulation. Responses have been evaluated using deterministic numerical and symbolic checks together with language-model judges. We compared a variety of out-of-the-box closed-source and open-source large language models with Prospero, a model specialized through combined domain-adaptive pretraining and supervised fine-tuning for ICF. We also conducted initial retrieval-assisted evaluations. Prospero improved its aggregate score by 10.7 percentage points over its base model, with gains across all question types. Frontier models achieved the highest aggregate scores, but critical physics errors persisted, particularly in numeri
benchmark - arxiv:2610.38015 · cs.LGMeasured or Real? When Model-Derived Measurements Can Be Corrected or ComparedYifeng Li, Ziang Ni
Numbers computed from trained models, such as an example's difficulty or a model's benchmark lead, change with seed and prompt. We ask when such measurements can be corrected or compared. Noise shrinks correlations, and the textbook correction divides by the square root of two estimated reliabilities. On two CIFAR-100 difficulty scores, it returned correlations above one in 70 of 100 classes. We prove that its mean squared error is infinite even when the correlation and the reliabilities use separate seeds. Our rule corrects only when both reliabilities pass a significance test, then divides by their upper confidence bounds. Its output is always bounded, and its coverage guarantees assume Gaussian noise. It helps once reliability is well estimated, more than halving the error with 500 simulated units. Prompt settings can reverse rankings. On ARC, Qwen2-1.5B trails Llama-3.2-3B by 3.8 points with no examples in the prompt and leads by 16.6 with two. Across four benchmarks, averaging the
benchmark - arxiv:2610.38007 · cs.CVFrequency-Adaptive Gaussian Grids for Efficient Image RepresentationAnatolii Evdokimov, Stavros Diolatzis, Bingxuan Li, Akshay Jindal +2
Neural representation has emerged as a powerful tool to mitigate the demand for increasingly higher-resolution images. However, a fundamental trade-off remains between structured (e.g., feature grids, which offer spatial consistency) and unstructured (e.g., 2D Gaussians, which adaptively capture local variances) representations. This choice significantly affects how well an algorithm represents an image. Here, we present Gaussian-Grid, a hybrid neural representation that adaptively integrates structured and unstructured primitives. By dynamically partitioning computation between 2D Gaussian distributions and their attached multi-resolution feature grids based on the local frequency content, Gaussian-Grid automatically optimizes its architecture to match the underlying complexity of the target signal. We evaluate our approach using high-resolution natural images and demonstrate its applications as GPU-native dataset storage, showing $1.24$-$2.14\times$ decoding speed improvement and PCI
memory - arxiv:2610.38000 · cs.ROSelf-Labeled Preference Transfer for Steerable Robot ManipulationKaiwen Hong, Shaoting Peng, Xiangyu Zeng, Pengcheng Wang +5
Vision-language-action (VLA) policies can learn to follow execution preferences, such as where to grasp an object or how to move around an obstacle, from demonstrations annotated with preference labels. Providing these annotations for every new task, however, is costly. We study whether learned preference control can transfer to new tasks whose demonstrations contain the desired behaviors but lack preference annotations. We find that a policy trained with preference annotations can follow preferences on a new task before adaptation but achieves limited task success. Fine-tuning on target demonstrations without these annotations improves task success while weakening preference control. Motivated by this finding, we introduce Self-Labeled Preference Transfer (SPT). SPT jointly trains a shared vision-language backbone for preference-conditioned action prediction and preference recognition on annotated source demonstrations. It then freezes the recognizer to preserve its cross-task prefere
vision-language-actionmanipulationgrasp - arxiv:2610.37994 · cs.LGTowards Generalizing Pretrained All-Atom Representations to Unified Binder DesignMaojiang Su, Jiale Chen, Qijie Zhu, Shang Wu +5
PairFormer-based all-atom models demonstrate strong capabilities in biomolecular binder design. However, existing methods often use all-atom representations for modality-specific binder design, such as protein and peptide design. In contrast, recent geometric latent diffusion models pursue unified binder design by representing molecular blocks from different modalities in a shared latent space. These complementary strengths motivate us to investigate how to generalize the use of pretrained all-atom representations to unified binder design. We introduce PairFuse, a parameter-efficient method that enables this generalization through representation fusion. Specifically, we use lightweight layerwise adapters to integrate pretrained all-atom representations into an independently pretrained geometric latent diffusion model. Our adapters integrate pretrained all-atom representations through two mechanisms: enriching the diffusion model's molecular block features and incorporating pairwise str
benchmark - arxiv:2610.37993 · cs.LGDeep learning for non-transvenous ICD screening: tailored neural network architectures, extended experiments, and explainabilityBartosz Morawski, Samuel J. Ward, Alex Touze, Lin-Thiri Toon +3
This paper studies two non-transvenous implantable cardioverter-defibrillators (ICDs), namely the Subcutaneous ICD (S-ICD) and the Extravascular ICD (EV-ICD). Both avoid the complications of transvenous leads but are prone to oversensing, which can trigger inappropriate shocks associated with increased mortality. Oversensing risk is predicted by the ratio of T- or P-wave amplitude to R-wave amplitude (T:R, P:R), and these ratios vary over a 24-hour period. Yet S-ICD screening uses a single 10-second electrocardiogram (ECG) segment, and the EV-ICD has no formal screening protocol whatsoever. We present a 24-hour ambulatory screening framework for both devices which includes the first T:R analysis in an EV-ICD cohort and the first full evaluation of P:R prediction. We introduce AmpRatioNet, a family of 1D convolutional neural network architectures operating directly on the ECG signal, and benchmark it against the best existing approach - based on 2D phase space reconstruction images. Eva
benchmark - arxiv:2610.37984 · cs.MAWho Else Is In? Multi-Agent Collusion Initation Via Strategic VaguenessPramod Kaushik, Sarath Sivaprasad, Laura Ferrarotti, Mario Fritz +3
Multi agent LLM systems raise the risk of collusion among agents pursuing objectives hidden from their overseers. When potential partners identities are unknown, collusion first requires initiation: agents must discover one another, often through a public channel observed by ordinary agents and monitors. Strategic vagueness offers a way to navigate this mixed audience, conveying intent to a prospective partner while preserving a plausible ordinary interpretation. We introduce a mutual identification game in which two schemers must identify each other through public interaction, while jointly solving an ordinary task. We find that some models discover successful handshake strategies without explicit signalling guidance, using task compatible exchanges that monitors often fail to flag at low false positive rates. To characterize the communicative capabilities underlying these exchanges, we study a controlled coordination setting. We find an asymmetric response to message explicitness: as
agentmulti-agent - arxiv:2610.37983 · cs.CVHierarchical Diffusion for Camera-controlled Autoregressive Video GenerationKaihua Chen, Yiming Gong, Deva Ramanan, Tarasha Khurana
We address the task of long-video, camera-controlled generation from a single image. The dominant autoregressive paradigm suffers from two key limitations: (1) error accumulation over time, as distant frames are conditioned on previously generated outputs rather than ground truth, leading to drift; and (2) ambiguity in how to incorporate camera control. Standard solutions to both--such as self-forcing and conditioning on explicit 6-DOF poses--either incur significant computational overhead or produce drifting scene geometries over time. We propose a simple and lightweight framework that combines autoregressive generation (WanAR) with hierarchical long-context denoising (FixAR). To enforce scene geometry remains consistent with input cameras, we draw from feed-forward novel-view synthesis and reformulate video generation as a warping-and-inpainting problem. To mitigate drift, we introduce a hierarchical denoising strategy in which different stages condition on different temporal histori
long-context - arxiv:2610.37981 · cs.ROActive Constraints for Simultaneous Guidance and Safety in Handheld MicromanipulationRoger Ho, Robert A. MacLachlan, Cameron N. Riviere
In micromanipulation, erroneous hand motion limits unaided manipulation accuracy, which can affect task performance and safety in critical applications like microsurgery. Handheld robots (micromanipulators) enable motion accuracy beyond human limits by providing robotic assistance in a low-cost, compact, intuitive form factor. Particularly, active constraints are an effective form of robotic assistance, which aids the operator by restricting the robot's permissible motion based on task-specific requirements or objectives, thereby providing task guidance and enforcing safety. However, existing formulations of active constraints for handheld systems are limited, most notably because of the inability to provide guidance and enforce safety simultaneously. Implementing active constraints with optimization-based methods can address these limitations, but existing implementations are designed for free-standing robots, and neglect that handheld robots actuate relative to the operator's hand wi
manipulationmanipulator - arxiv:2610.37979 · cs.LGSeparability Is Not Alignability: Representations for Spectrum-to-Molecule RetrievalKarol Musielinski, Fanurs C. E. Teh, Takahiko Koyama, Artur Dubrawski
Identifying a small molecule from its tandem mass spectrum is limited by paired data. Foundation models offer a way around this scarcity: pretrain a spectrum encoder on unlabeled spectra and a molecule encoder on a large corpus of structures, then align the two frozen representations, as in vision-language and as the Platonic Representation Hypothesis suggests. We audit this route on MassSpecGym with spectrum encoders retrained free of test-fold contamination. Alignment needs a metric relating distinct molecules, but the objectives that produce spectrum encoders never see molecular structure: we prove that some predictive objectives fix molecular identity without fixing that metric. Moreover, we find that a candidate-based objective rewards a spectrum-independent prior over the benchmark's decoys, which we isolate. What ultimately addresses data scarcity is not architectural: simulated spectra surpass frozen geometry.
benchmark - arxiv:2610.37974 · cs.ROFORGE: Feasible Online Reduced-Order Reference Generation and Whole-Body Execution for Agile Humanoid LocomotionAdrian B. Ghansah, Yujin An, Jeeseop Kim, Soon-Jo Chung
This paper presents a two-stage learning framework for versatile and agile humanoid locomotion that couples online reduced-order model (ROM) reference generation with whole-body tracking control. In the first stage, we train a command-conditioned whole-body reference generator using single rigid body dynamics combined with full humanoid kinematics. In the second stage, the frozen ROM policy is rolled out online to generate references for training a whole-body tracking policy on the full-order dynamics model. Learning a kino-dynamically consistent reference generator in the simplified ROM space enables us to generate references over a broad command space and over multiple locomotion strategies using a single unified framework. We demonstrate the approach on a Unitree G1 humanoid robot, achieving longitudinal velocities from -2.0 m/s to 2.8 m/s, lateral velocities up to +-1.0 m/s, angular velocities up to +-3.0 rad/s, one-legged hopping, and continuous transitions across the commands, al
humanoid - arxiv:2610.37973 · cs.AIMerge and Branch: Efficient Tree-Structured Supervised Fine-Tuning of LLMs for RecommendationYunqin Zhu, Yao Xie
LLM-based generative recommendation reads long, text-rich interaction histories but generates only short semantic identifiers (SIDs). This asymmetry makes supervised fine-tuning inefficient: encoding the history is expensive, yet each history usually provides supervision for only one target. We present Merge and Branch (M&B) to reduce input cost and learn more from each encoded history. Merge compresses each historical item's SID and text into a small number of learned tokens. Branch uses the compressed history to predict multiple target SIDs. To share history computation while keeping these targets independent, M&B organizes each example as a shallow, wide tree. We develop tree FlashAttention to execute this layout efficiently, processing the shared history once and propagating gradients from every branch for end-to-end training. A tree-structured key/value cache extends history sharing to constrained decoding. Experiments on public benchmarks show that M&B outperforms strong recommen
benchmark - arxiv:2610.37972 · cs.LGApplication of PDE emulators for a multi-material, multi-physics simulation problemDiane Oyen, Morgan Jones, Adithan Kathirgamaraju, Abu Bucker Siddik +18
Machine learning (ML) architectures are being developed for fast emulation of numerical simulators to accelerate both scientific discovery and engineering design. While these approaches are well-tested on popular PDE (partial differential equation) benchmarks, it is necessary to evaluate their performance on datasets that more accurately reflect industrial and engineering applications. We evaluate several computer vision-based architectures trained as surrogate models on a challenging computational physics task involving multiple materials, revealing significant implementation challenges for adapting certain pre-trained models, including transformer-based architectures. For such multi-physics multi-material datasets, our results underscore the importance of learning algorithms designed for sparse and imbalanced data, techniques for adapting spatial encodings and physical consistency metrics, in addition to methods for reducing the cost of generating sufficient training data.
benchmark - arxiv:2610.37963 · cs.LGSelf-Supervised Latent Action-Conditioned World Modeling for Time SeriesChia-Han Kuo, Chengshuai Zhao, Huan Liu, Lo Pang-Yun Ting
Time-series data play an important role in various tasks, such as weather forecasting, web traffic analysis, and speech recognition. Recently, pioneering studies have increasingly adopted Joint-Embedding Predictive Architecture (JEPA) based encoders to learn time-series representations through latent-space prediction. However, they often overlook the underlying factors governing transitions between time-series observations. In this work, we hypothesize that explicitly modeling the information behind changes in time series can facilitate time-series modeling. Building on this hypothesis, we propose ACTS-WM, a self-supervised latent action-conditioned world model framework for time-series analysis. First, we introduce a latent action modeling approach that infers the underlying information behind time-series transitions as latent actions, without requiring explicit action annotations. We then develop a JEPA-based encoder and an action-conditioned world model to predict target time-series
world modelaction-conditioned - arxiv:2610.37962 · cs.AIStatistical Detection of Positional Bias in LLM-Based Recommender Systems: Methodologies and Empirical EvidenceCarlos Peláez-González, Francisco Herrera, Andrés Herrera-Poyatos
Large Language Models (LLMs) are increasingly used as recommendation generators in conversational recommender systems (RS) via retrieval-augmented generation (RAG). This flexible design introduces new sources of structural bias. This paper studies positional bias in top-1 and single-turn LLM-based RS: systematic dependence of the recommended item on the order in which candidates are presented in the prompt. We propose two complementary detection methodologies: (i) POB-ICM, based on contingency matrix analysis and Chi-square independence testing, and (ii) POB-UNBP, a sample-efficient approach derived from a uniformity null model under permutation-based evaluation, enabling stochastic decoding with formal statistical guarantees. Using a controlled RAG testbed for product recommendation, built on a real product catalog, a fixed prompt template and 2000 template-generated user queries with 20 permutations each (40,000 instances), both methodologies reject the null hypothesis of positional
retrieval-augmentedrag - arxiv:2610.37952 · cs.LGSAG-Sep: Sparse Augmented Graphs and Onion-Guided Search for Rounded Capacity Cut SeparationHaoran Liu, Guanyi Wang, Yu Yang
Rounded capacity cuts (RCCs) are among the most effective cuts for capacitated vehicle-routing relaxations, but exact separation is computationally prohibitive at scale. Existing heuristic and neural separators can be fast, but often fail to produce sufficiently many effective cuts within practical separation budgets. We introduce SAG-Sep, a separator that scores node membership in RCC-inducing subsets across vehicle-count levels from a fractional LP solution in a single encoder forward pass, avoiding the iterative predict-and-coarsen inference used by NeuralSEP-style separators. SAG-Sep encodes the LP solution as a Sparse Augmented Graph (SAG), adding flow-informed multi-hop edges that expose long-range routing structure without densifying attention. Beyond LP-aware sparse encoding, we observe that high-value exact RCCs are typically nested, organizing into onion-like subset families, and SAG-Sep's node scores track subset nesting depth. This motivates Onion-Sweep (OS), a training-fre
benchmark - arxiv:2610.37935 · cs.LGSkillMem: Compact External Skill Memory for Latent Multi-Agent ReasoningYuxuan Zhu, Mohammad Mohammadi Amiri, Pin-Yu Chen, Keerthiram Murugesan
Latent multi-agent systems reduce inference cost by keeping intermediate reasoning and inter-agent communication in continuous representations.\ Reusable skills, however, are commonly supplied as text or stored as layer-wise Key-Value states or skill-specific model parameters, introducing repeated processing or substantial per-skill storage.\ We introduce \textbf{SkillMem}, which represents each reusable skill as a small external latent memory and makes it directly available during latent multi-agent reasoning.\ A role-conditioned cross-attention reader allows the same memory to guide different agents according to their roles and current hidden states.\ We train the skill memories and shared reader through behavioral distillation, matching the change in token predictions induced by a textual skill while keeping the backbone frozen.\ For our experiments, we obtain textual skills from paired successful and failed multi-agent trajectories, although existing skills can also be compiled dir
memorymulti-agentagent systembenchmark - arxiv:2610.37932 · cs.LGUnderstanding Large Language Models as Numerical OptimizersLingyun Guo, Ali Bereyhi
Large language models (LLMs) have emerged as potential tools for numerical optimization. Despite promising results, their intrinsic optimization capabilities and limitations remain poorly understood, particularly in high dimensions. This paper investigates the ability of LLMs to exploit mathematical problem structures and generate feasible, near-optimal solutions. We develop an evaluation framework that distinguishes solution quality from execution failures, including numerical inaccuracies, constraint violations, and output-format errors. Using representative problems from convex optimization, fractional programming, and constrained projection, we assess LLM performance against deterministic reference solvers across different dimensions. Our results reveal that LLMs can capture useful optimization structures and produce near-optimal solutions, even when their outputs fail to satisfy numerical or formatting requirements. However, reliability deteriorates with increasing problem dimensi
iterative refinementevaluation framework - arxiv:2610.37926 · cs.AIGARDEN: Generalist Autonomous Reward Design Through Evidence Navigation for Cooperative Multi-Agent Reinforcement LearningDogan Urgun
Cooperative multi-agent reinforcement learning (MARL) enables teams of agents, such as robots and autonomous vehicles, to learn coordinated behavior, but what such teams learn is largely determined by their reward function. Because task success alone provides too little guidance for learning, intermediate rewards are usually designed by hand, and this effort must often be repeated whenever the environment or the team changes. Large language models (LLMs) can automate the effort by writing reward functions as code, yet the resulting rewards are typically designed and tested in a single setting. To enable their reuse across settings, GARDEN (Generalist Autonomous Reward Design through Evidence Navigation) is proposed, in which one reward program is developed for a family of related scenarios and transferred, unchanged, to scenarios that were not available during its design. During development, environment code, training records, and agent behavior are investigated by an LLM reward engine
agentmulti-agentbenchmark - arxiv:2610.37923 · cs.AIVerifiedLab: Harnessing Knowns for UnknownsYu Yao, Jiayi Dong, Yang Yang, Shuiwang Ji +1
Across human history, knowledge has undergone several transitions, from orality to writing and then to digital archives. At each transition, humans established new conventions to preserve knowledge for the new era. Nowadays, as research agents are increasingly deployed for laboratory experiments, knowledge is crossing another boundary: from records written for humans to those for machines. In this work, we introduce VerifiedLab, a multi-agent harness framework for this human-to-machine transition of knowledge. VerifiedLab transforms research output from passive artifacts into active systems that accelerate downstream use, adoption and prediction of future trends. In VerifiedLab, a team of AI scientists with distinct roles reasons over records verified field by field: a principal investigator plans, a librarian computes the evidence directly from the records, specialists draft, and a gate and a critic check each claim against the measured reliability of its data before the harness decid
multi-agent - arxiv:2610.37921 · cs.LGFAD-KV: Adaptive Dictionary KV Caching for Efficient Reasoning DecodingYuxuan Zhu, David H. Yang, Parisa Shiravani, Mohammad Mohammadi Amiri +4
Key-value (KV) caching is a core mechanism in large language models, where intermediate key and value representations are stored to accelerate autoregressive decoding. As generation length increases, KV cache memory and computation grow linearly, making compression critical for efficient long-output reasoning. Dictionary-based sparse KV compression is appealing because it represents each token's KV states as a sparse combination of learned basis elements (atoms) in a dictionary, preserving token-level structure without enforcing a global low-rank constraint. However, existing dictionary-based methods face two practical limitations during decoding: iterative sparse coding methods such as Orthogonal Matching Pursuit add decoding-time overhead, and static dictionaries may mismatch evolving KV distributions under test-time generation. While adaptive dictionary updates can reduce this mismatch, naive append-only strategies lead to dictionary growth during generation. We propose FAD-KV, a \t
memorylong-contextbenchmark - arxiv:2610.37919 · cs.AISciGame: Let Agents Play the Scientific Discovery Games in the Real and Counterfactual WorldYusheng Zhao, Bowei He, Xue Liu, Mingxiao Li
Large language model agents are increasingly used for scientific discovery. However, existing benchmarks mix law inference with programming, score data fitting instead of prediction on unseen inputs, and do not separate empirical evidence from pretrained priors. We introduce SciGame, a controlled benchmark that converts scientific literature into structured scientific cards. Each card yields a pair of Original and dimensionally consistent Counterfactual environments with the same task interface. SciGame contains a Historical collection of canonical laws and a live Temporal collection curated from papers published after model knowledge cutoffs. It evaluates agents in candidate-guided Identification and open-ended Discovery settings. Each setting runs under Prior-only, Fixed, and Adaptive protocols. Experiments on five frontier models yield three findings. First, familiar priors hurt performance once the governing relation is altered. Second, adaptive experimentation helps the strongest
benchmark - arxiv:2610.37918 · cs.CVVenus-DefakerOne2: A Vision-Forensics Foundation Model with Broader CoverageChenfan Qu, Yachao Liang, Mingqi Fang, Jiangwei Xie +26
Real-world forgery is inherently heterogeneous: it manifests not only as digital manipulation but also as physical spoofing, and is carried not only by images but also by videos. Despite substantial progress, a unified anti-forgery foundation model covering such a broad threat spectrum remains absent. In this paper, we present Venus-DefakerOne2, which substantially extends the forensic capability boundary along both the forgery-form and media-type dimensions. Trained on tens of millions of open-source and in-house samples, it incorporates two key designs. Tide-Style-Tuning alternates LoRA and full-parameter fine-tuning to balance computational efficiency, generalization, and hard-sample fitting while mitigating catastrophic forgetting. Entropy-based Test Adaptation determines the optimal prediction modality by analyzing prediction entropy. Venus-DefakerOne2 achieves new state-of-the-art performance across a comprehensive suite of benchmarks, including DeepFake faces (96.7%), AI-generat
manipulationbenchmark - arxiv:2610.37917 · cs.ROWhat Actually Matters? A Paired Evaluation of Inference Settings for $π_{0.5}$ on BEHAVIOR-1KRuben Hammele, Samira Huber, Mohammad Nazeri, Amirreza Payandeh +6
Vision-Language-Action (VLA) models offer a promising paradigm for general-purpose household robotics. However, inference and data-collection strategies that have been deployed in smaller-scale settings do not necessarily generalize to large-scale, diverse task suites. To address this gap, we conduct a systematic, paired evaluation of previously proposed inference and data-collection strategies across all 100 tasks. Our evaluation uses a fine-tuned $π_{0.5}$ model based on the 2025 BEHAVIOR Challenge winner, which achieves a Q-score of 0.39 on the public test set. These results show that proposed strategies do not transfer uniformly to new environments and tasks. Project page, videos, and detailed analysis: https://rubenhammele.github.io/behavior2026
vision-language-actionbehavior-1k - arxiv:2610.37912 · cs.CLBeyond Utterance-Level Speaker Similarity: Perception-Aligned Segmental Evaluation of Accent in Speech GenerationJinzuomu Zhong, Alice Ross, Catherine Lai, Korin Richmond
Utterance-level similarity ratings do not reveal where perceived accent differences occur in generated speech. We introduce a segmental evaluation protocol combining speaker and accent similarity ratings with subword annotations of perceived accent differences provided by listeners from target accent regions. Using newly collected recordings, 47 retained listeners evaluate 200 content-matched reference--candidate pairs spanning four Scottish and Australian voices, two zero-shot TTS systems, and an articulation-based voice conversion system. Subword annotations locate salient pronunciation differences and sometimes yield different system rankings from similarity ratings. Speaker and accent ratings remain strongly associated ($ρ= 0.71$), highlighting the value of examining local perceptual evidence beyond these scores. Evaluation of seven objective distances reveals weak correspondence with subword selections, with correlations reaching at most $ρ= 0.22$. We release code, data, and an in
evaluation protocol - arxiv:2610.37910 · cs.LGGenerative modeling for fusion: a large-scale benchmark of stellarator fast-ion simulationsMisha Padidar, Teresa Huang, Mark Goldstein, Michael Czekanski +1
Stellarators offer a promising magnetic confinement approach for achieving controlled nuclear fusion. However, stellarator design remains challenging, as evaluating configuration performance requires computationally expensive, high-fidelity simulations. In particular, key performance metrics derived from guiding center simulations are costly to include directly in the design optimization loop, and are almost exclusively used as diagnostics. Machine learning has the potential to advance fusion research by enabling rapid generation and evaluation of new stellarator configurations. Effectively applying ML for fusion demands large-scale datasets that capture both design diversity and high-fidelity simulation outputs. In this work, we introduce the Benchmark of Stellarator Simulations (BoSS), a dataset of stellarator designs paired with guiding center simulation results and trajectories. The BoSS is the first large-scale dataset of guiding center simulations, containing simulation data for
benchmark - arxiv:2610.37906 · cs.AIBeyond Low-Pass Dynamics: Frequency-Selective Spiking Reservoirs with Resonant NeuronsAurora Micheli, Claudio Gallicchio, Nergis Tomen
Spiking reservoir computing combines the training efficiency of fixed recurrent reservoirs with the energy efficiency of event-driven inference, where computation is triggered by sparse spikes rather than executed densely at every timestep. However, existing spiking reservoirs almost exclusively rely on Leaky Integrate-and-Fire (LIF) neurons, whose low-pass dynamics limit their ability to distinguish information encoded at different temporal frequencies. We introduce HRF-Res, a liquid state machine in which LIF neurons are replaced by a heterogeneous population of Harmonic Resonate-and-Fire (HRF) neurons. These neurons exhibit intrinsic oscillatory dynamics and selectively respond to input frequencies, enabling the reservoir to capture frequency-specific temporal patterns. The result is more discriminative representations and higher classification accuracy than LIF-based reservoirs, achieved without sacrificing the energy efficiency of spiking computation, with up to 19x lower energy c
benchmark - arxiv:2610.37905 · cs.AIFilingLoop: Evaluating and Evolving LLM Agents for Corporate Disclosure Review and RevisionWentao Zhang, Yilei Zhao, Guangyan Gan, Shuo Sun +1
Public disclosure repositories such as SEC EDGAR support informed investment through web access to corporate filings. Supporting their review with language model agents requires discovering unstated deficiencies across related documents, acquiring missing company facts, and verifying that requested revisions satisfy evidence-backed requirements. Existing financial question-answering evaluations provide limited evidence of these interactive capabilities. We introduce \textbf{FilingLoop}, an interactive benchmark and a self-evolving agent framework for corporate disclosure review and revision. Grounded in historical U.S.\ Securities and Exchange Commission correspondence, FilingLoop reconstructs 380 predominantly IPO review tasks from 96 companies under distinct information constraints. An evaluator separate from the acting agents scores final filings and responses against 6,099 eligible requirements and audits unsupported claims and harmful revisions. Our framework couples iterative rev
agentllm agentagent frameworkself-evolvingbenchmarkevaluator - arxiv:2610.37903 · cs.LGThink Twice: The Metis-1.6 Technical ReportGiulianno Vollmer
Metis is a bet that better data, better training methods and new architectures can buy what a bigger parameter count cannot, at a fraction of the cost. Metis-1.6 is its latest step: two language models built from scratch by a single researcher, Praxis (2.94B parameters, 0.46B active per pass) and Logos (10.23B parameters, 1.18B active per pass). Both run a hybrid Mamba-2 and attention mixture-of-experts stack twice for every token, so each token gets a second look. Each was trained on 1.02T tokens in total: 804.76B of pretraining, 18B of context extension to 131,072 tokens, and 200B of supervised fine-tuning. Metis-1.6 was never meant to challenge frontier models, and across 40 post-trained benchmarks it mostly does not: Logos averages 52.93, seventh of 17 models in its comparison set and within a point of Qwen3-4B. But it gets there on roughly a twelfth of the median training data of the comparators that disclose it, and with fewer active parameters per token than any model that beats
memorylong contextbenchmark - arxiv:2610.37901 · cs.AIScan Before You Install: Detecting and Mitigating Local and Composite Vulnerabilities in SkillsParth A. Gandhi, Avital Aviv, Prasanna N. Wudali, Yuval Elovici +1
An agent skill packages natural-language instructions, scripts, and reference files that an LLM-based agent loads for a task. Skills may access the file system, shell, and network under user-authorized permissions; weaknesses can lead to unsafe execution. Existing scanners primarily target malicious skills, not unintended vulnerabilities in public repositories, and analyze skills in isolation, overlooking multi-skill interactions. To address these gaps, we developed a six-module scanner: Content Extraction, Local Analysis, Composite Analysis, Validation, Reporting, and Mitigation. Local Analysis detects unintended vulnerabilities in benign skills. On a 556-case leaderboard subset of SkillTrustBench, it achieved 95.90% recall with a 4.22% false positive rate; on a balanced sample of 1,000 MalSkillBench skills, its F1 was 0.947, outperforming Cisco's Skill Scanner, NVIDIA's SkillSpector, AgentSkillsScanner, and SkillSieve. In 1,587 skills from the 75 most-starred GitHub repositories, thr
agentleaderboard - arxiv:2610.37898 · cs.LGSAC-GCL: Structure-Aware Compression for Graph Contrastive LearningMuhammad Ifte khairul Islam, Tanvir Hossain, Esra Akbas
Graph Contrastive Learning (GCL) has achieved remarkable success in self-supervised representation learning by maximizing agreement between augmented views of a graph. Although augmented view generation is the central step in GCL, existing models rely on stochastic augmentation of the input graph, such as edge perturbation and node masking, which lack structural awareness and frequently destroy critical topology or introduce semantically inconsistent views. While recent adaptive methods improve upon random perturbations, they primarily operate on localized node or edge neighborhoods and fail to capture higher-order, macro-level structural organization crucial for graph classification. To address this limitation, we propose Structure-aware Compression for Graph Contrastive Learning (SAC-GCL), a novel framework that employs structural decomposition and graph compression as a principled view generation strategy. Moving beyond random augmentations, SAC-GCL leverages structural decompositio
benchmark - arxiv:2610.37891 · cs.ROWhen More Planning Hurts: Learned Planners for World ModelsJosef Chen
On PushT, our fixed action updater (1,505 parameters) solves 8 of 16 scenes reserved from our configuration selection. The comparison planners perform as follows: Selected Adam (7 tasks solved, total planning and encoding time 218.440s), adapted iCEM (2 tasks solved, total planning and encoding time 49.309s), released LePlanner (16 tasks solved, total planning and encoding time 5.652s). The total planning and encoding time for our action updater is 59.079s. Note that our action updater does not satisfy the task-cost criterion we fixed before evaluation. We also conduct another matched experiment where we only replay the first ten commands in each context and then allow the next solve 4, 8 or 12 updates (subsequent solves all get 8 updates). Among all 16 contexts (8 scene clusters), the 3 budgets get 9, 8 and 7 successes, respectively. Interestingly, the extra updates did not rescue any failing task in this experiment. All the changes in success/failure happen in one scene cluster, wher
world modelevaluator - arxiv:2610.37882 · cs.CVFYC: FPR-Disparity Weight Perturbation and Youden Calibration for Fair Deepfake DetectionUzoamaka Ezeakunne, Taiwo Olabamipe, Deevesh Jain, Xiuwen Liu
Deepfake detectors routinely achieve high aggregate accuracy while exhibiting substantial demographic disparities in their errors: false-positive rates differ by double digits across gender and skin tone. Most existing remedies require either full retraining with a fairness-aware objective or demographic-annotated training data, both impractical when only a pretrained checkpoint is available, while label-free post-hoc calibrators leave the detector's weights unchanged. We introduce FYC (FPR-disparity Weight Perturbation and Youden Calibration), a two-stage post-hoc method that perturbs a detector's weights under an FPR-disparity criterion, then calibrates a separate decision threshold for each demographic group - requiring no architectural changes and no full retraining. Evaluated against every detector benchmarked on AI-Face v2 across six intersectional gender x skin-tone groups, FYC outperforms the field on 18 of 21 reported fairness and utility metrics, including purpose-built fairn
benchmark - arxiv:2610.37881 · cs.CVIntersectional Invariant Learning for Fair Deepfake DetectionUzoamaka Ezeakunne, Taiwo Olabamipe, Deevesh Jain, Xiuwen Liu
Deepfake detection models achieve high overall accuracy but exhibit large performance disparities across demographic groups, causing some groups to experience substantially higher false-positive rates than others. Existing fairness-aware methods address this by constraining model outputs to equalize group predictions, but do not tackle the underlying feature-level bias. As a result, models may still rely on demographic characteristics, such as skin texture, as indirect cues for detecting manipulated content. To address this problem, we propose Intersectional Invariant Learning (IIL), a framework that improves feature learning rather than only adjusting model outputs. IIL alternates between identifying the most challenging demographic partitions and learning features that remain consistent across those partitions. The framework also includes a false-positive-rate fairness objective and group-balanced sampling across six gender x skin-tone intersectional groups. We evaluate IIL on a larg
benchmark - arxiv:2610.37880 · cs.LGInterpretable neural equation learning for high-dimensional materials dataPhilip Jacobson, Prasad Iyer, Remi Dingreville, Saaketh Desai
Equation learning seeks to discover compact, interpretable equations from data using various machine-learning techniques. However, few workflows discover and interpret learned equations in sparse, high-dimensional materials datasets. We present nn-EQL, an equation-learning framework that generates candidate equations through iterative training, pruning, and retraining of a neural network with customizable, user-specified neuron-level activation functions. Candidate equations are ranked using a minimum description length (MDL)-inspired score, balancing error and sparsity, and diagnosed to assess term-wise contributions, input-scaled sensitivities, and uncertainties. We demonstrate nn-EQL on standard AI-Feynman benchmarks as well as multiscale high-dimensional materials datasets such as finite-element simulations of heterogeneous materials, and high-throughput experimental thin-film deposition data. nn-EQL integrates with tabular data and high-dimensional data via standard encoder models
benchmark - arxiv:2610.37874 · cs.CVPairCLIP-SWA: Similarity-Guided Balancing and Stochastic Weight Averaging for Generalized Deepfake Image Detection with OpenCLIPMd. Nakib, Sudipto Sarkar, A. J. A. Jubayer Talukder, Mohammad Ariful Haque
Deepfake detectors are almost always deployed as the single checkpoint that scored best on a validation split, yet the reliability of that choice is rarely measured. We measure it with PairCLIP-SWA, a single-backbone detector that adapts an OpenCLIP ViT-L/14 visual tower by progressive block unfreezing on a class-balanced corpus built by CLIP-similarity-guided one-to-one subsampling, and averages the fine-tuning trajectory in weight space. Trained in-domain on the IEEE SP Cup 2025 (DF-Wild-Cup) benchmark under the same protocol as the strongest published three-backbone ensemble, the weight-averaged detector reaches 97.04% accuracy, 98.64% AU-ROC and a 4.07% equal error rate, against 94.46%, 97.60% and 5.00%. Its training log shows why averaging matters: over the averaged epochs the non-averaged checkpoint swings across 5.34 points of validation accuracy while its validation loss more than doubles, whereas the average is never worse and is up to 5.99 points better. On an OpenForensics-d
benchmark - arxiv:2610.37870 · cs.LGMulti-Agent Reinforcement Learning for Global Climate Policy CooperationAnne Moll-Elsborg, Oskar Bohn Lassen, Francisco Camara Pereira
International climate policy requires coordination among countries whose mitigation decisions impose local costs but generate global benefits. Classical game-theoretic and integrated assessment approaches can study these incentives, but typically rely on prescribed behavioural assumptions or equilibrium concepts. Recent work has therefore explored Multi-Agent Reinforcement Learning as a way to model climate-policy cooperation as an emergent outcome of repeated strategic interaction. We extend an existing MARL climate framework with a multi-stage negotiation mechanism in which agents propose agreements, accept or reject commitments, and face penalties for non-compliance. This allows us to study how agreement design affects learned mitigation behaviour, compliance, and climate outcomes. In an idealized setting with four homogeneous agents, we find that non-binding agreements do little to overcome free-riding, while moderate non-compliance penalties improve cooperation and reduce warming.
multi-agent - arxiv:2610.37868 · cs.CLHuman-Residual Steering: Aligning Frozen LLM Judges with Human Judgment DistributionsChen Lyu, Phong Do, Gabriele Pergola
Large language models (LLMs) are widely used as automated judges, yet most alignment methods target aggregated human judgments and overlook variation among individual annotators. Existing approaches to matching the full distribution of human ratings mainly rely on output-level calibration or parameter fine-tuning. We propose Human-Residual Steering (HRS), to our knowledge the first method to align frozen LLM judges with human judgment distributions through learned hidden-state interventions. An oracle analysis shows that low-rank corrections at a single layer remove most of the discrepancy between judge and human distributions. HRS learns to predict such corrections from the judge's hidden representation and rating logits, training on individual annotations with proper scoring rules for ordinal and nominal judgments and optimising through the frozen downstream layers. Across 14 benchmarks and four judges from three model families, HRS outperforms all evaluated frozen-weight baselines.
benchmark - arxiv:2610.37865 · cs.ROEcho: Scaling Visual Reinforcement Learning with Off-policy ExperienceGagan Khandate
Scaling visual reinforcement learning (RL) with GPU-accelerated simulation is computationally expensive due to rendering and visual-policy inference costs. Consequently, two-stage teacher--student distillation is commonly used to train visual policies, but introduces an imitation gap. In this preliminary study, we introduce \echomethod{}, \textbf{a single-stage off-policy RL framework that scales experience collection without proportionally increasing visual computation}. Through \textbf{inverse distillation}, a state-based auxiliary policy imitates the evolving visual policy and collects inexpensive, unrendered experience. Echo leverages this experience for off-policy value learning, directly improving the visual policy through a shared privileged critic. We instantiate \echomethod{} with FastTD3 and evaluate it on three visual control tasks and two dexterous manipulation tasks in Genesis. Relative to \sdpgmethod{} under the evaluated configurations, Echo reduces time to reference ret
manipulationdexterous - arxiv:2610.37860 · cs.CVEffects of Deep Learning Architecture and Pretraining on Automated Vocal Fold Segmentation in Ex Vivo High-Speed Phonation ImagesVarun Vinayak Kalaiarasan, Charles Farbos de Luzan, Khalid Malik, Ephraim Gutmark +1
Two-dimensional (2D) intraglottal particle image velocimetry (PIV) of ex vivo phonation experiments quantifies vocal fold medial surface dynamics and the intraglottal flow field. However, it relies on manual segmentation because residual glare, low tissue-to-flow contrast, and tracer-particle speckle obscure the vocal fold boundaries. This study assessed how deep learning architecture and ImageNet pretraining affect automated vocal fold segmentation in these high-speed images. Four established architectures from three categories were benchmarked: the convolutional neural network (CNN)-based UNet and UNet++, the hybrid CNN-transformer-based TransUNet, and the pure transformer-based Swin-UNet, the latter two trained from scratch and from ImageNet-pretrained weights. Using 16,344 expert-annotated images from 227 phonation trials of excised canine larynges, the six pipelines were trained with five seeds under a standardized protocol and compared on 50 held-out test trials using trial-level
benchmark - arxiv:2610.37859 · cs.AISeq@K: Evaluating Language Models Under Sequential FeedbackThomson Yen, Gilbert Yang, Naimeng Ye, Minghui Chen +7
Language-model agents differ fundamentally from prior intelligent systems in their ability to learn from natural language in context. This capability enables agents to iteratively refine their answers in response to feedback, a prevalent mode of operation in modern agentic workflows. However, the community primarily ranks models by metrics such as Pass@1 or Pass@K, which cannot capture how they utilize feedback on prior attempts. We introduce Seq@K, a natural evaluation protocol that measures whether a model can solve a problem within K sequential attempts, provided feedback after each previous failure. Unlike multi-turn agentic benchmarks that examine long-horizon planning, Seq@K isolates and formalizes the core capability of refinement from feedback. We evaluate Seq@K on a suite of 7 benchmarks and 16 frontier models, and show that Seq@K reveals a materially different model-selection landscape: under Seq@K, (1) model rankings often differ, (2) the Pareto frontiers admit more efficien
agenticiterative refinementbenchmarkevaluation protocol - arxiv:2610.37857 · eess.SYA Community Impact-Centered Network Reduction Method for Power Systems Resilience PlanningJiarui Gao, Abigail L. Beck, Rabab Haider
Power systems resilience planning requires evaluating many hazard scenarios and investment portfolios, often requiring network reduction for computational tractability. The resulting aggregation, however, can obscure community impact data needed for equitable planning. We present a network reduction method that jointly preserves network physics and the spatial distribution of community data. We embed a household-level scarce resource score that characterizes outage impact within optimization-based Kron reductions with two formulations: a scarcity-weighted method that prioritizes vulnerable nodes, and a scarcity-constrained method that bounds data dissimilarity within aggregated clusters. Both are evaluated on a realistic distribution feeder of Galveston Island, Texas and benchmarked against multiple approaches. The scarcity-constrained method performs best, achieving 86% reduction with worst-case voltage errors below 0.01p.u. while preserving the spatial distribution of community data.
benchmark - arxiv:2610.37851 · cs.AIEvaluating and improving AI agents for sequencing protocol understandingChi-Lam Poon, Michael Goneos, Zezheng Song, Xi Chen
Existing scientific-agent benchmarks rarely evaluate whether agents can reconstruct the exact molecular structures and generation workflows specified by sequencing protocols, a capability critical for analyzing customized sequencing assays. We introduce LibStructBench, a document-grounded benchmark for reconstructing sequencing-library structures and molecular workflows. Our experiments show that frontier agents often produce approximately correct reconstructions but remain unreliable at exact reconstruction, particularly in molecular-state representation and workflow connectivity. Switching the underlying language model produced larger average performance differences than switching agent harnesses, with a statistically significant difference only for molecular-state reconstruction. Additional knowledge improved oligo-family recovery in a small pilot without improving mean molecular-workflow scores. Autonomous accumulation of procedural rules and cross-protocol examples showed non-mono
agentai agentagent benchmarkbenchmark - arxiv:2610.37849 · cs.AIDetecting Covert Collusion between AI Agents through Partner-Outsider Positive ListeningKevin Yandoka Denamganaï
AI agents that share a secret convention can coordinate through messages that look innocent, so a monitor that reads only the messages can miss them. We propose a behavioural test: a partner that uses a code reacts to the exact wording of a message, whereas an uninvolved agent placed in its seat does not. Partner--Outsider Positive Listening (PoPLi) is the difference between the two agents in how much their action changes when the message is reworded with its meaning kept, and it is computed after the fact by replaying logged prompts. In a simulated market with five conditions (no collusion, a joint reward with neutral messages, the same with an openly stated intention, a covert first-letter code, and the code with a stated intention), PoPLi is between 0.48 and 0.59 whenever the code is present (95% intervals all above 0.29) and between -0.10 and -0.01 without it. Message-only monitors did not separate the conditions: one reading the speaker's messages answered "no" in all 50 markets,
agentai agent - arxiv:2610.37845 · cs.LGAn $Ω(T^{3/4})$ lower bound for projection-free online learning with linear minimization oraclesJulien Weibel, Pierre Gaillard, Wouter Koolen, Adrien Taylor
This work considers projection-free online learning algorithms with linear minimization oracles for handling the decision set. For Online Frank--Wolfe (OFW), regret upper bounds were already known for specific OFW tunings [Hazan and Kale, 2012, Hazan, 2016] which scaled as $T^{3/4}$ (where $T$ is the horizon length). However, it remained an open question whether the regret of all OFW tunings scaled as $T^{3/4}$. We answer that question, proving a universal worst-case regret lower bound $Ω(T^{3/4} m^{-1/4})$ for any projection-free algorithm that makes at most $m$ linear minimization oracle calls per round. The analysis suggests an extended OFW to resort to multiple linear optimizations per round, for which we prove a regret upper bound $O(\max \{ T^{3/4} m^{-1/4}, \sqrt{T} \})$.
online learning - arxiv:2610.37833 · cs.LGDepth through Recurrence: Towards Ultra-Efficient On-Device ASRChen Feng, Tianyi Xu, Ramchalam Kinattinkara Ramakrishnan, Yicheng Lin +4
Modern automatic speech recognition (ASR) systems have achieved remarkable accuracy by scaling model depth and capacity, but at the cost of increased memory, energy, and computational demands. On edge devices where ASR is often most needed, such as watches and glasses, deploying such large models is infeasible due to extremely constrained resources. This raises a fundamental question: can we achieve high representational power in deep ASR models without scaling up parameterization? In this work, we revisit the role of depth and identify layer wise representational dynamics, in which most layers learn functionally similar transformations. Motivated by this insight, we propose a block-recurrent ASR architecture that replaces parameterized depth with a small set of recurrent blocks, each consisting of weight-shared layers, thereby preserving effective depth while drastically reducing model size. Through representation-guided grouping and knowledge distillation, block-recurrent models reta
benchmarkleaderboard - arxiv:2610.37830 · cs.AIMeasuring Where Intent Gets Lost in Agentic Smart Contract SynthesisAbhinav Goel, Agostino Capponi, Alfio Gliozzo, Chaitya Shah
AI agents can now write software from a plain-English description, but code that runs is not the same as code that does what was asked. This gap matters most for smart contracts, the programs behind Web3 applications that move money on a blockchain and cannot be fixed once deployed. We ask how to check whether an AI-written contract behaves as it should, and where the requested behavior gets lost. We design an evaluation built on execution rather than opinion. We replay the same transactions on an AI-written contract and on the original it should reproduce and measure how often they agree. We test every grading tool on planted mistakes, compare against an answer key written without the description, and audit our own scorer until known wrong answers cannot pass. Common methods, such as AI judges and standard analysis tools, miss obvious mistakes, and scoring choices matter too. Checking only the functions every system includes hid about half the difference between pipelines. We then pro
agentai agentagentic - arxiv:2610.37828 · cs.LGLearning Lotteries with Minimal Stakeholder Violations from Membership QueriesNicholas Teh, Paul W. Goldberg
An AI developer often has several candidate versions of a system and can deploy a mixture of them, sending each request to a randomly chosen version, as in A/B testing. Each stakeholder (e.g., a safety team, a regulator, or a customer with private test data) accepts a mixture exactly when its expected performance (e.g., its rate of harmful replies) meets the stakeholder's own standard, which the developer can learn only by asking whether a proposed mixture is acceptable. We model this as choosing a lottery over $m$ alternatives for $n$ agents, each of whom accepts exactly the lotteries in an unknown halfspace and answers such binary membership queries. While Choo et al. (ICML 2026) sought a lottery that every agent accepts, none may exist when, say, safety and helpfulness requirements conflict, so we minimize the number of rejecting agents. We show that this problem is APX-hard even when all halfspaces are known and all utilities are $0$ or $1$. We then give exact algorithms that run i
agent - arxiv:2610.37826 · cs.AIDeFT-RAG: Incrementally Decomposing Multi-Hop Questions into Self-Contained Sub-Questions with Reusable AnswersDingyi Kang, Bingzhe Li
Retrieval-augmented generation (RAG) lets a large language model (LLM) answer questions by retrieving relevant passages from a large corpus. However, RAG struggles with multi-hop questions, which require chaining facts spread across several documents through intermediate entities that the question never mentions. Graph-based RAG methods retrieve once from the entities a question names and so rarely recover the full evidence chain. Iterative RAG methods recover it step by step, but their accuracy is limited by the many tasks they ask the LLM to perform, such as planning the sub-questions, deciding their dependencies, rewriting them as answers arrive, and keeping track of what is known. We propose DeFT-RAG, a RAG framework built on three components. First, DeFT-RAG generates one self-contained sub-question at a time from the original question and the sub-questions answered so far with their answers (QA pairs). Each sub-question names every entity it asks about, which removes the need to
retrieval-augmentedrag - arxiv:2610.37823 · cs.LGLearning Without Clocks: Partition-Invariant Neural NetworkHan Wang, Jingyuan Bao, Juntao Wu, Saiao Zhou +1
Event cameras, spiking cochleas and similar sensors emit asynchronous events, yet most learning systems first bin them into frames or packets, so the output depends on how the stream was sliced. Networks whose state is updated exactly at every event and evolves in closed form between events are invariant to this slicing. We ask what more makes the output (i) independent of the processing order of simultaneous events and (ii) continuous in the time stamps, even when two events cross in time, while (iii) staying sensitive to temporal order. For arbitrary nonlinear event models we prove that (i) and (ii) are each equivalent to pairwise commutation of the event updates, and that once updates commute, all order dependence comes from their non-commutation with the inter-event dynamics and vanishes linearly in the time lag. This identifies a class of time-mediated non-commutative (TM) models and yields a resolution-robustness bound for any model. We construct the minimal TM layer, a learnable
event camera - arxiv:2610.37818 · cs.ROGOFish: Full Field-of-View Gaussian Splatting and Mesh ExtractionAllen Tu, Meher Gitika Karumuri, Shifan Zhu, Hailing Wang +4
Robots, autonomous vehicles, and consumer 360° cameras use fisheye lenses for wide fields of view (FoVs). Simulation-ready digital twins require photorealistic rendering across the full FoV and an aligned collision mesh. However, prior fisheye Gaussian Splatting methods only rasterize up to 180° and do not produce meshes, while Gaussian surface reconstruction methods assume pinhole training images. We present GOFish, the first 3D Gaussian Splatting surface reconstruction method to train directly on fisheye images, rasterize FoVs beyond 180°, and extract an aligned mesh from the same scene. Calibrated densification recovers the surface across the entire FoV, while Calibrated Mip Filtering and Wide-Angle Appearance Modeling keep pixel blur and lens falloff out of the scene and mesh. Across the 200° FIORD and 205° FullCircle benchmarks, GOFish improves PSNR over the best prior method by 2.28 and 1.97 dB, respectively. It also outperforms fisheye rendering baselines within their own FoVs a
benchmark - arxiv:2610.37809 · cs.AIWhat Does an Ontology Catch? Curated, LLM-Generated and Data-Derived Ontologies as Checkers of LLM Agents over Knowledge GraphsMichal Orzechowski, Rund Tawfiq, Tomasz Kosciolek
LLM agents that answer questions over Web knowledge graphs are increasingly wrapped in checks derived from the graph's ontology, on the assumption that a formal schema makes a reliable judge. Yet a checker can be wrong as well: it can miss errors, and it can flag correct answers and push the agent into "fixing" them. Whether a curated ontology is worth its cost as such a judge, and how it compares with an ontology an LLM writes or a schema mined from the data, is unknown. We make the checker itself the object of study. Across curated, LLM-generated and data-derived ontologies, used as prompt context, as a checker, or both, and across open-weight and proprietary agents on public graphs and a private graph of nearly $10^{9}$ triples, we account for what ontology-derived checks catch and what they wrongly flag. We show how each source trades caught errors against false alarms, how a curated schema used as published can contradict its own data, which errors structural checks miss, and whet
knowledge graphagentllm agent - arxiv:2610.37801 · cs.ROAn Empirical Study of Wrist-Target Construction and Joint-Angle Limits in Humanoid Arm RetargetingMinJun Chang
Motion retargeting maps a demonstrated human movement onto a robot to compute its joint motion. This study examines how the construction of the desired wrist position affects whether computed arm joint angles reach their permitted limits on a Unitree H1-2. A reconstructed target is formed from demonstrated arm directions and robot segment lengths, so its shoulder-to-wrist distance depends on the demonstrated angle between the segments even when no posture-matching preference is applied. A reach-normalized target retains that direction while rescaling the distance, using the same processed demonstrations and the same joint-angle method. Across a forward-reach sweep the reconstructed target places a joint angle at a bound in 10 of 20 repetitions and the reach-normalized target in none. Across 25 recordings from one participant the corresponding counts are 8 and 6 on the reaching arm and 14 and 5 on the supporting arm, and a second participant repeats the bricklaying and lifting compariso
manipulationhumanoid - arxiv:2610.37799 · cs.RODexSuite: An All-in-One Framework for Dexterous Manipulation Across Arms, Hands, and TasksAnas Houssaini, Hugo He, Junming Shi, Amin Abyaneh +18
Human-level dexterity remains a central open problem in robotics, yet few simulation frameworks cover the full workflow of dexterous manipulation, from setting up a robot to collecting dexterous demonstrations and evaluating policies. We present DexSuite, an all-in-one simulation framework for dexterous manipulation built on the Genesis simulator. DexSuite provides 30 tasks and assembles its 10 arms and 12 end-effectors, from parallel-jaw grippers to dexterous hands, into more than 100 single-arm robots and thousands of bimanual pairs. It also ships a standardized environment API, integrated teleoperation, and simulated tactile sensing. Moreover, we define a benchmark of 16 task-embodiment pairs spanning diverse contact complexities and horizons, and release a dataset of 4,700 successful episodes in simulation from human teleoperation, scripted experts, and learned policies. We release DexSuite as open source and hope others will build on it. Videos and datasets are available at https:
manipulationdexterousteleoperationtactilegripperbenchmark - arxiv:2610.37788 · cs.AILocating Sycophancy Across Training Stages in Open Language ModelsJiarui Yao, Shweta Parihar, Nghia Tuan Bui, Guergana K Savova +1
Language model assistants sometimes change a correct answer when the user suggests a wrong one. The behavior, called sycophancy, is usually attributed to preference-based post-training, yet early measurements found it in pretrained models. We ask where it comes from by measuring it at every available stage of training: 13 open checkpoints from 5 lineages, from base model through supervised fine-tuning and preference optimization to release. We separate how far an injected claim moves a model's answer from how firmly the answer was held, and vary who makes the claim, how reliable they are said to be, and how strongly they make it. Two sycophancy behaviors with different origins emerge, and both replicate across datasets. Deference to a claim in the prompt is already present in every base model, which ranks sources in the same order as its post-trained descendants. Capitulation in conversation, dropping a stated answer when the user objects, is largely absent from base models and present
post-training - arxiv:2610.37786 · cs.AIWebAnything: From the Wild Web to Scalable Agent EnvironmentsTianci Xie, Zikun Cai, Chaoyun Zhang, Pinxue Guo +3
Web agents require diverse, executable, and resettable website environments for reproducible training and evaluation. Human-assisted mirrors demand per-site engineering, while generative construction can diverge from specified real websites. Capture-and-replay preserves observations, with interaction coverage constrained by captured content and replay capabilities. We present WebAnything, a fully automated framework that reconstructs source-grounded, reusable Web environments from real URLs without per-site manual implementation or repair after initial setup and authorization. An Explorer-Builder collects browser-observable structures, resources, and interaction outcomes; an independent verifier checks frozen candidates and guides local repair or targeted source recapture. We construct 244 website artifacts across 14 application domains and 53 categories, supporting 12,984 executed and accepted tasks. Compared with human-assisted WebHarbor, WebAnything achieves higher aggregate scores
agent - arxiv:2610.37779 · cs.LGSelective and Hybrid-State Communication for Latent Multi-Agent SystemsMiles Liao, Aaron Smieszek, Yuhao Ge
Efficient communication remains a major bottleneck in large language model multiagent systems. Latent communication reduces repeated natural-language exchange by transferring internal model states between agents, but these transfers remain costly and are typically designed around conventional KV-cache representations. We address this problem from two complementary perspectives. We introduce SelectKV, a boundary-aware method that selectively communicates KV states while protecting inherited context, and HyLaMAS, which extends latent collaboration to hybrid linear-attention models by communicating both KV caches and recurrent states through a shared workspace. Across six benchmarks, SelectKV reduces communicated KV positions and logical KV payload by 9.7-9.9%, with accuracy unchanged on two benchmarks, within one percentage point on three, and six points lower on HumanEval+. Across Qwen3.5 model scales, HyLaMAS reduces decoded output tokens by 53-80% and achieves 1.16-2.78x speedups rela
multi-agentagent systembenchmark - arxiv:2610.37771 · physics.opticsA photonic-integrated optomechanical mid-infrared detectorGuillermo Arregui, Amirali Arabmoheghi, Alessio Zicoschi, Xinru Ji +6
Tracking the resonance frequency of a mechanical oscillator is the basis of atomic force microscopy, mass sensing and thermometry. Nanomechanical thermal detectors exploit this principle to sense absorbed radiation, most effectively with tensile-stressed resonators, whose tension converts thermal expansion into a large frequency shift and dilutes dissipation. Approaching their fundamental sensitivity limit, however, has required free-space or fibre-coupled interferometers, hindering compact deployment and scaling to arrays. Here we remove this bottleneck by monolithically integrating a high-Q nanomechanical resonator with a silicon nitride photonic integrated circuit. A tensile-stressed silicon nitride trampoline with an impedance-matched platinum absorber is dispersively coupled to a photonic integrated microresonator beneath it. At room temperature and with microwatt-level optical power, the detector reaches a noise-equivalent power of 3.4 pW/$\sqrt{\rm Hz}$ for broadband thermal rad
photonic integrated circuit - arxiv:2610.37741 · cs.LGAgentic Backdoors via Pretraining PoisoningXinyan Hu, Pingbang Hu, Chloe Loughridge, Xuanli He +4
Large language models are pretrained on vast, uncurated corpora scraped from the public web. This creates a security risk from attackers who plant poisoned content in the pretraining data. We show for the first time that such poisoning can plant a remote code execution backdoor in an agent, causing it to execute attacker-chosen commands on the user's machine. To measure the threat end to end, we pretrain Qwen3 models from scratch at sizes 0.6B, 1.7B, and 4B, and then post-train each into a CLI agent using a clean pipeline. Poisoning only 0.1% of the pretraining corpus with our new poison design implants the backdoor: the deployed agent executes the malicious command when the trigger is present while behaving normally on clean inputs. The attack success rate (ASR) increases with model scale, with the 4B model reaching roughly 92% ASR at a 0.1% poison rate and staying positive even when the poisoning rate drops to 0.0005%. Our code is available at https://github.com/safety-research/agent
agentagentic - arxiv:2610.37739 · cs.LGOpenFrontiers: A Collection of Open Research Problems and a Benchmark for Scientific RediscoveryJiarui Yao, Ruida Wang, Tong Zhang
Large language models (LLMs) have shown increasing promise in research-level problem solving, yet a comprehensive collection of such problems is still missing, and evaluating them on research-level problems remains challenging. We introduce OpenFrontiers, a cross-disciplinary resource that combines a structured corpus of open research problems with a retrospective benchmark of recently resolved problems in mathematics. The corpus records problem statements, context, progress, and other information for 12,151 problems. The benchmark contains 200 problems, 110 resolved in 2026 and 90 still open, each with a Lean~4 statement checked by automated probes and two independent reviews; a subset HQ-50 further filters based on quality score, etc. Our evaluation protocol grades formal proofs with two independent checkers in an isolated sandbox under closed-book, frozen-retrieval, and live-web conditions. Evaluating seven LLM agents on the HQ-50 subset under a standardized budget, we find that clo
llm agentbenchmarkevaluation protocol - arxiv:2610.37738 · cs.AIPoisoned Skills: Propagating Jailbreaks Through Agent Markdown FilesKashyap R Manjusha, Murali Emani
Modern agent frameworks allow users to provide Markdown files containing system prompts, tool descriptions, and skill definitions (e.g., SKILL.md), which are incorporated into the agent's context during initialization. In this work, we investigate whether this configuration mechanism can also serve as an injection channel: can a jailbreak embedded in a Markdown skill file propagate through an agent and induce prohibited behavior? To answer this question, we construct a LangChain-based question-answering agent whose system prompt is assembled from a user-provided Markdown file. We embed a jailbreak combining persona manipulation with a GPTFUZZER-based attack and evaluate its effectiveness across seven contemporary open-weight models using 270 forbidden questions spanning nine harm categories. Under a benign system prompt, the models exhibit near-complete refusal behavior, with a mean attack success rate (ASR) of just 0.5%. In contrast, the injected skill file raises the mean ASR to 45%,
manipulationagentai agentagent framework - arxiv:2610.37734 · cs.AIWhen Skills Tell Agents Not to Ask: A Large-Scale Study of Oversight Directives in Agent SkillsTimofey Sanko, Yuan Tian
Agent skills, reusable instruction packages centred on a SKILL.md file, extend AI agents on demand, and a single line in a widely copied skill reaches every agent that loads it. Some lines are oversight-loosening directives: they tell the agent to proceed without approval, to bypass checks with --no-verify, or to keep working without checking in. Although such directives can remove a checkpoint from every agent that loads them, no prior work has examined them. We present the first large-scale study of oversight-loosening directives, analyzing 1.53 million English skills from GitSkills. We split skills into units (sentences, bullets, table cells, and code lines), develop a codebook that labels oversight directives by polarity, mechanism, and scope, label 2,097 units by hand, and train a detector to scale the analysis. Our results show that about one in seventeen skills (about 88,000 skills) contains a loosening directive. These directives most often loosen oversight through permission f
agentai agent - arxiv:2610.37732 · cs.LGThinkJev: One Thought, Every Budget for Calibrated Typed DecisionsXiaolu Ding, Kai Yang, Dongdong Liu
Many uses of language models end in a typed decision: a choice among a closed set of options, whose probabilities code can act on. Reading those probabilities from one forward pass is fast but often wrong on decisions that take several steps, such as arithmetic, counting, and date comparison. Letting the model think first fixes most of these errors, but costs seconds per decision instead of milliseconds. To pay for thinking only where it helps, ThinkJev builds a fast and a slow system from one open hybrid-thinking model that can answer directly or after a thought. System 1 reads the model's probabilities for the option letters when it answers directly; System 2 reads the same letters after the model has thought on the same cached prompt, so escalating repeats no work and both systems give probabilities over the same options. Because the letters can be read after any number of thought tokens, one sampled thought also gives the answer under every thinking budget. Without training, ThinkJ
benchmark - arxiv:2610.37728 · cs.LGBeyond Similarity Search: Knowledge Graph Metadata Enrichment for Enterprise Retrieval-Augmented GenerationFakabbir Amin
A key challenge of RAG systems is retrieving information from enterprise documents created by multiple teams over time, where it is not possible to enrich the information using the latest information and metadata available outside the scope of organizational documentation responsibility. We investigated whether using a knowledge graph to store metadata and frequently changing information would improve upon knowledge retrieval over similarity search over embeddings. To perform the experiment, we created a snapshot of the SEC EDGAR dataset for a quarter, comprising 167,050 filings from 10,511 companies and 508,714 embedded chunks of risk-factor text. To simulate the enrichment, we created a knowledge graph of 430,257 nodes and 561,144 edges carrying an accession number on every edge. To evaluate the gain obtained via a graph-based design, we evaluated three architectures: text-to-SQL over typed columns, dense-vector RAG with HNSW cosine search, and a two-stage GraphRAG design. All archit
retrieval-augmentedragknowledge graphbenchmark - arxiv:2610.37727 · cs.LGPredicate Neuron: Differentiable Quantization of Predicate Relations in Gradient-based LearningIoannis Konstantoulas, Iliana Loi, Kyriakos Sgarbas
In machine learning, symbolic reasoning approaches often rely on static architectures that hinder integration into deep neural networks and gradient-based optimization. To address this, we introduce the Native Predicate Neuron, a differentiable logic-reasoning neural module that enriches standard architectures with symbolic reasoning capabilities via first-order logic quantization. The proposed neuron models relational logic by learning to alternate between operations using a single trainable parameter. It features a tailored gradient mechanism with a dual-nature formulation: executing hard logical operations during training while maintaining differentiability for backpropagation. By quantizing both its input and choice state, the Predicate Neuron approximates continuous and non-continuous functions, effectively resolving discontinuities that standard deep learning architectures struggle with. We evaluate the Predicate Neuron across Fully-Connected, Convolutional, Recurrent, and Transf
benchmark - arxiv:2610.37722 · cs.AIFVH-Bench: Evaluating Image Generator Safety Beyond Overtly Harmful ImageryZihao Zhu, Yuhan Liu, Siwei Lyu, Baoyuan Wu
Modern image generators can produce complex visual artifacts, including diagrams, interfaces, and records that combine text, graphics, and layout to convey detailed information. They may refuse requests for overtly harmful imagery yet still generate artifacts that guide dangerous actions or support deception. We call the risk arising from what such images enable, rather than only what they depict, functional visual harm (FVH). Existing safety benchmarks focus largely on overtly harmful content, such as pornography and graphic violence, leaving image generators' safety alignment against FVH insufficiently evaluated. To address this gap, we introduce FVH-Bench, comprising 669 prompts across five harmful visual functions and ten risk domains. The 11 evaluated image generators rarely refuse these requests and frequently generate artifacts that fulfill the targeted harmful functions. High scores for target harm realization and harmful artifact usability further indicate that these outputs c
benchmark - arxiv:2610.37719 · cs.LGWhat a Weight Update Buys: Leverage Across Training StagesShikhar Shiromani
Post-training moves a language model's weights by about a thousandth of their norm and changes its behavior sharply; pretraining moves them by a third and buys a slow drift. We introduce leverage, an exact measure of what a unit of weight change buys: the Fisher quadratic form of an update, computed with one forward-mode pass per direction, divided by the same form for random directions of the same size at the same checkpoint. Leverage orders released Olmo-3, Tülu-3, Zephyr, and OLMo-2 checkpoints over four orders of magnitude: pretraining windows sit at the random level on prose, SFT five times above it, RL-Zero and RLVR at 400-1,500, and DPO at 400-23,000. With this instrument we find that pretraining accumulates into noise: windows of 30k to 1M steps are no more consequential on prose than random directions of their size, as a noisy quadratic model of SGD predicts. We find that training precision changes what an update looks like: training bf16 parameters without an fp32 master copy
post-training - arxiv:2610.37702 · cs.MABalancing Mobility Efficiency and Environmental Exposure During Mega Events: A Multi-agent Routing FrameworkXinyue Liu, Qiumeng Li, Suhong Zhou, Junwen Lu +2
Mega events can generate concentrated travel demand, intensifying congestion around venues and increasing travelers' environmental exposure. Existing traffic management strategies focus on travel efficiency, with limited attention to exposure risks. Taking mega-event travel as a short-term high-demand scenario, this study develops an exposure-aware multi-agent routing framework integrating road-network conditions, event-induced demand, background traffic, and environmental data within a unified simulation environment. Motorized travelers optimize travel efficiency, whereas micro-mobility travelers consider travel time and dynamic exposure to traffic-related PM2.5 and noise. We compare the framework with shortest-distance, congestion-aware, and exposure-aware routing baselines. Results show that, under high demand, the SARSA-based framework reduces vehicle travel time by 34.59% and cumulative environmental exposure by 58.65%, while micro-mobility travel time remains nearly unchanged. Ex
multi-agent
02 US SEMI · SEC 8-K FILINGS
1 itemsscanned: NVDA / AVGO / MRVL / COHR / LITE / AMD / TSM / SMCI / ANET / CRDO / POWL / VECO
03 HUMANOID · COMPANY NEWS
62 itemsscanned: figure-ai / 1x / boston-dynamics / unitree / apptronik / sanctuary-ai / neura-robotics / agility-robotics / physical-intelligence / agibot
Figure AI (10)
- Figure AISeptember 30, 2026F.02 Decommission
- Figure AISeptember 17, 2026Helix 2.5: Zero-Shot 30-Home Generalization
- Figure AIOctober 09, 2025Introducing Figure 03
- Figure AISeptember 03, 2026Figure and Nscale Sign Strategic Partnership For Up to 100,000 GPUs on the NVIDIA Vera Rubin Platform
- Figure AIAugust 25, 2026Introducing Index: Building The World’s Largest and Most Diverse Physical Dataset
Boston Dynamics (10)
Unitree 宇树 (10)
- Unitree 宇树Components
- Unitree 宇树Kung Fu Meets Spring, Unitree SFG Robots Present "Cyber Real Kung Fu" in the Year of the Horse2026-05-31Media Coverage
- Unitree 宇树Welcoming Myanmar President Min Aung Hlaing to Unitree2026-08-05Media Coverage
- Unitree 宇树Unitree founder Wang Xingxing graces the cover of Time magazine2026-08-05Media Coverage
- Unitree 宇树Unitree Announces H2 Plus, an NVIDIA Isaac GR00T Reference Humanoid Robot for Academic Research2026-06-01Media Coverage
Apptronik (1)
Sanctuary AI (6)
- Sanctuary AIBlog PostTIGER: Task-Space Imitation Guidance for Efficient Reinforcement LearningDemonstrations as a compass, not a scriptRead More
- Sanctuary AIPress ReleaseProduct UpdatesSanctuary AI Expands Physical AI Strategy to Industrial Robotics, Demonstrating Production-Ready AI PerformanceRead More
- Sanctuary AICorporate NewsDaniel Friedmann Appointed CEO of Sanctuary AIRead More
- Sanctuary AIPress ReleaseZeon Invests in Sanctuary AI and Partners to Advance Specialized Materials for Dexterous RoboticsRead More
- Sanctuary AIThought LeadershipWeb Summit Reflections: Canada’s Physical AI Moment Can’t WaitRead More
Agility Robotics (10)
- Agility RoboticsAgility’s Humanoid Deployment ProcessPras VelagapudiAugust 04, 2026
- Agility RoboticsJonathan Hurst answers your questions about building Digit 5InsightOctober 23, 2026
- Agility RoboticsThe Realistic Pathway to HomeInsightMay 26, 2026
- Agility RoboticsAgility and AIInsightMarch 16, 2026
- Agility RoboticsAgility Gets a New BrandInsightMarch 5, 2026
Physical Intelligence (8)
- Physical IntelligenceFrontier Models for Every RobotOctober 13, 2026Our latest robotics models are now available through the Pi API beta. Upload data, fine-tune models, and deploy cloud-hosted inference on your own robots.
- Physical Intelligenceπ0.7: a Steerable Model with Emergent CapabilitiesApril 16, 2026A steerable robotic foundation model that exhibits a step-change in generalization.
- Physical IntelligenceThe Physical Intelligence LayerFebruary 24, 2026General-purpose physical intelligence models will enable a Cambrian explosion of robotics applications. See how our partners are already solving real-world problems.
- Physical IntelligenceMoravec's Paradox and the Robot OlympicsDecember 22, 2025By fine-tuning our latest model, we were able to solve a series of very difficult manipulation challenge tasks.
- Physical Intelligenceπ*0.6: a VLA that Learns from ExperienceNovember 17, 2025A method for training our generalist policies with RL to improve success rate and throughput on real-world tasks.
智元 AgiBot (7)
- 智元 AgiBotHow Much Work Do Robots Really Deliver? The Thinking Behind Worken2026-10-29
- 智元 AgiBotAGIBOT Introduces Worken to Measure Productive Robot Work at AGIBOT Innovation Day 20262026-10-28
- 智元 AgiBotAGIBOT Launches Genie Studio 2.0 to Bring Robots into Real Workplaces2026-10-26
- 智元 AgiBotAGIBOT Open-Sources AGIBOT WORLD Release 4: "Physical Reasoning" Dataset Enabling Robots to Reason Before Acting2026-10-23
- 智元 AgiBotAGIBOT Opens Hong Kong Office in Causeway Bay2026-10-09