AI news archive.
19,864 articles, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchRationale Distribution Distillation via Flow Matching for Distributional Continuous Latent Reasoning
arXiv:2610.19871v1 Announce Type: cross Abstract: Sampling multiple chain-of-thought rationales from a language model can uncover correct solutions missed by a single attempt. However, these rationales are typically leng…
- arXiv cs.AIresearchNeural Wasserstein Splines: Representation-Invariant Second-Order Priors for Trajectory Inference
arXiv:2610.19878v1 Announce Type: cross Abstract: Population dynamics are often observed only as unpaired snapshots, which leave the evolution between measurements underdetermined. Second-order regularization can constra…
- arXiv cs.AIresearchFrustratingly Easy Majority Voting Using Length Conditioning
arXiv:2610.19890v1 Announce Type: cross Abstract: Majority Voting (MV) is a simple and widely used strategy which improves the reasoning accuracy of Large Language Models (LLMs) by allocating additional inference-time co…
- arXiv cs.AIresearchDeepPV-MRF: End-to-End Partial-Volume Fraction Estimation from MR Fingerprinting Time Series
arXiv:2610.19894v1 Announce Type: cross Abstract: Background: Estimating subvoxel tissue fractions from magnetic resonance fingerprinting (MRF) is challenging. Conventional dictionary matching and segmentation pipelines…
- arXiv cs.AIresearchUniversal Sample-Based Embeddings of Higher-Order Objects
arXiv:2610.19899v1 Announce Type: cross Abstract: Neural networks operate naturally on finite-dimensional vectors, yet many problems require reasoning about higher-order objects such as distributions, functions, and grap…
- arXiv cs.AIresearchMultimodal Medical Image Parsing with Task-Specific Low-Rank Adapters
arXiv:2610.19901v1 Announce Type: cross Abstract: Adapting a single vision-language model to medical image parsing requires training choices that reflect the demands of each task. We study task-specific adaptation with Q…
- arXiv cs.AIresearchMUCO: A Mani-Unit-Structured Manipulation Framework for Compositional Long-Horizon Robot Tasks
arXiv:2610.19913v1 Announce Type: cross Abstract: Learning and executing long-horizon robot manipulation as complete task trajectories raises data-collection costs, lim?its capability reuse, and causes local failures to…
- arXiv cs.AIresearchDoes Machine Translation Retain Media Bias?
arXiv:2610.19945v1 Announce Type: cross Abstract: Within the most recent two decades, machine translation (MT) has been advanced from a research area to a mature, deployed technology for prose text and most language pair…
- arXiv cs.AIresearchLearning Graph Energies for Finite-Time Generation
arXiv:2610.19946v1 Announce Type: cross Abstract: Energy-based models can represent distributions over discrete structures such as molecular graphs, but an energy that is accurate at equilibrium need not yield accurate s…
- arXiv cs.AIresearchQuiet Triggers, Loud Footprints: A Tri-Space Measurement Study of Backdoor Stealthiness
arXiv:2610.19947v1 Announce Type: cross Abstract: Backdoor attacks on deep neural networks are commonly evaluated by attack success rate and benign accuracy, yet their stealthiness remains inconsistently defined and meas…
- arXiv cs.AIagentsAtomSearch: Evidence-Grounded Tree Search for Agent Skill Optimization
arXiv:2610.19951v1 Announce Type: cross Abstract: Agent skills adapt frozen models through editable external instructions. Revising several rules at once makes changes difficult to inspect, while judging each local edit…
- arXiv cs.AIresearchSafe-FCL: Continual Risk Awareness and Fine-Grained Safety Aggregation for Federated Continual Learning of LLMs
arXiv:2610.19968v1 Announce Type: cross Abstract: Federated fine-tuning enables large language models (LLMs) to learn collaboratively from distributed client data, but malicious client updates can compromise global model…
- arXiv cs.AIresearchEmpirical Boundaries of Reservoir Computing for Few-Shot Sequence Pattern Tasks
arXiv:2610.19997v1 Announce Type: cross Abstract: Efficient sequence modeling remains a fundamental challenge in machine learning. Despite the success of Transformers, recurrent neural networks (RNNs) have recently re-em…
- arXiv cs.AIresearchA Strong Harness Narrows the Model Gap: A Comparative Evaluation for Legislative Document-to-XML Conversion
arXiv:2610.19999v1 Announce Type: cross Abstract: Legal-content reproduction requires matching the source exactly in section hierarchy, verbatim text, citations, and amendment spans; a dropped citation or mislabeled amen…
- arXiv cs.AIresearchProvable Transformer Context Compression for In-Context Reinforcement Learning
arXiv:2610.20004v1 Announce Type: cross Abstract: Transformers can perform in-context reinforcement learning without parameter updates. However, existing theoretical constructions often require context sizes that grow wi…
- arXiv cs.AIresearchOn the Convergence of Semi-Gradient TD: A Mean Field Perspective
arXiv:2610.20014v1 Announce Type: cross Abstract: Semi-gradient temporal-difference learning is a core algorithmic component of reinforcement learning, but bootstrapping with function approximation generally induces a no…
- arXiv cs.AIresearchMorphAct: Robot Policies Morphing with the Worlds They Act In
arXiv:2610.20019v1 Announce Type: cross Abstract: Progress toward general intelligence in the physical world requires scaling robot learning across diverse environments and interactions. However, existing robot policies…
- arXiv cs.AIresearchSWE-Nexus: Automated Composition of Software Engineering Tasks from Test Execution
arXiv:2610.20038v2 Announce Type: replace-cross Abstract: Historical code changes provide convenient task boundaries, but do not identify which production functions selected tests exercise or support composing validated…
- arXiv cs.AIagentsCoRedSkill: Stealthy Skill Composition Attacks via AST-Guided Injection
arXiv:2610.20049v1 Announce Type: cross Abstract: Agents extend their capabilities through reusable skills that package instructions, executable code, and resources. Composing these skills enables complex task workflows,…
- arXiv cs.AIagentsGeoCloak: An Agentic Framework for Semantics-Preserving Geolocation Privacy via Peripheral Clue Manipulation
arXiv:2610.20055v1 Announce Type: cross Abstract: The rapid advancement of large vision language models (LVLMs) has made geolocation inference from images highly accurate and accessible. This poses serious privacy risks:…
- arXiv cs.AIresearchFoveaVLA: Foveation Actions Replace the Wrist Camera in VLAs
arXiv:2610.20064v2 Announce Type: replace-cross Abstract: We present FoveaVLA, a Vision-Language-Action (VLA) model that replaces the wrist camera with learned foveation actions. Wrist cameras provide close-up detail but…
- arXiv cs.AIresearchWhich Knowledge Should LLMs Trust? Uncovering a Generalizable Steering Direction Across Knowledge Conflicts
arXiv:2610.20067v1 Announce Type: cross Abstract: Large Language Models (LLMs) frequently encounter conflicts between internal parametric knowledge and external contextual knowledge, yet the mechanisms governing this sel…
- arXiv cs.AIresearchTrainable at Depth, Solvable at Small Size: What the Loss Surface of a Polynomial Activation Dictates for Designing, Initializing and Training Networks
arXiv:2610.20069v1 Announce Type: cross Abstract: What makes a deep network train, and what shape is its loss surface? We answer both with one construction at two scales: a polynomial activation f (z) = z + {\lambda}z 3…
- arXiv cs.AIresearchRethinking Prediction Horizons: Decoupling Model Queries from State Updates in Few-Step MeanFlow
arXiv:2610.20071v1 Announce Type: cross Abstract: MeanFlow predicts interval-average velocities, but conventional few-step sampling ties the interval queried from the model to the duration of the state update. We separat…
- arXiv cs.AIresearchAFCS: Atomic-Fact-Conditioned Synthesis for Factual Knowledge Pretraining
arXiv:2610.20086v1 Announce Type: cross Abstract: Synthetic data is increasingly used in language-model pretraining. Existing knowledge-corpus synthesis pipelines, including WRAP and Nemotron-CC, usually take a source pa…
- arXiv cs.AIagentsTask-State Governance: A Virtual Machine for Verifiable and Recoverable GUI Agent Execution
arXiv:2610.20100v1 Announce Type: cross Abstract: Graphical user interface (GUI) agents now execute workflows across real applications, but when execution drifts, users must determine what changed, what should have remai…
- arXiv cs.AIresearchFrom Spectral to Frobenius: When Muon Loses Its Spectral Implicit Bias
arXiv:2610.20112v1 Announce Type: cross Abstract: Prior work on implicit bias shows that gradient descent selects the Frobenius max-margin direction on linearly separable problems, whereas Fan et al. (2025) established s…
- arXiv cs.AIresearchACR-Net: Mitigating Semantic Dominance via Contrastive Acoustic-Semantic Decoupling
arXiv:2610.20114v1 Announce Type: cross Abstract: Speech Emotion Recognition (SER) models and Audio LLMs fail when vocal tone contradicts textual semantics. Extensive training on congruent data engenders a rigid semantic…
- arXiv cs.AIresearchDreamHMR: From Video Generation to Human Motion Modeling via Structural Focusing
arXiv:2610.20124v1 Announce Type: cross Abstract: Human mesh recovery (HMR) provides a geometric foundation for understanding human motion, but accurately reconstructing both body motion and fine hand articulation from v…
- arXiv cs.AIresearchMultiPartyTurnBench: Benchmarking Floor Control in Multi-Party Spoken Dialogue
arXiv:2610.20172v1 Announce Type: cross Abstract: Full-duplex speech systems must decide when to enter, hold, or relinquish the conversational floor, even as multiple speakers overlap or backchannel without taking it. Ex…
- arXiv cs.AIresearchCross-Model In-Context Neurofeedback Reveals (Limited) Activation Control in LLMs
arXiv:2610.20179v2 Announce Type: replace-cross Abstract: Probes on activation vectors are widely used to detect misbehavior in language models, but their success relies on the assumption that models cannot control what…
- arXiv cs.AIagentsSWE-Stream: Scaling Executable Training Data for SWE Agents via Progressive Construction and Elastic Streaming
arXiv:2610.20193v1 Announce Type: cross Abstract: Software engineering agents have advanced rapidly, but scaling their training remains constrained by the limited availability of high-quality executable tasks. Developmen…
- arXiv cs.AIagentsWhat Makes a Social Agent's Interactions Effective? Learning from Verifiable Outcomes in Paired Worlds
arXiv:2610.20195v1 Announce Type: cross Abstract: Effective social interaction requires more than fluent responses: agents must reach plans that fit the participants' needs and circumstances. Existing evaluations of soci…
- arXiv cs.AIagentsWhat and Where in Identifying and Quantifying Conventions in Humans and AI Agents
arXiv:2610.20205v1 Announce Type: cross Abstract: Cooperative success rests in part on implicit communication. Partners converge on conventions: regularities that give an action meaning beyond its direct effect. Yet conv…
- arXiv cs.AIresearchScaling Multilingual Safety and Reasoning Evaluation through Reasoning-Guided Translation
arXiv:2610.20221v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed across languages, yet their safety and reasoning capabilities remain predominantly evaluated in English. Translatin…
- arXiv cs.AIagentsPAWA: Panoramic Action-conditioned World Modeling for Multiple Agents
arXiv:2610.20240v1 Announce Type: cross Abstract: A multiplayer world model must predict how the independent actions of several players change a shared environment, including changes outside a player's current field of v…
- arXiv cs.AIresearchRetrieval Over Training: Similarity search-based Model Selection for Time Series Anomaly Detection
arXiv:2610.20243v2 Announce Type: replace-cross Abstract: Anomaly detection is a fundamental task for time-series analytics, with important implications for the downstream performance of many applications. Despite the la…
- arXiv cs.AIagentsP2Patch: Exploit-Guided Vulnerability Repair
arXiv:2610.20245v1 Announce Type: cross Abstract: Large language models are increasingly deployed as autonomous agents for automated program repair, patching software vulnerabilities through multi-step reasoning and tool…
- arXiv cs.AIagentsLong-Horizon RL Needs Muon
arXiv:2610.20256v1 Announce Type: cross Abstract: Long-horizon agents must coordinate reasoning, tool use, and environment feedback over extended interactions. Learning from sparse outcome rewards requires weighting traj…
- arXiv cs.AIagentsHarness Policy Distillation: Aligning Agent Policies with Runtime Semantics
arXiv:2610.20261v1 Announce Type: cross Abstract: Deploying a general-purpose model as an agent means running it inside a product-specific runtime harness that decides which actions are executable, how state advances, an…
- arXiv cs.AIresearchFUSE: Fairness-aware Update Screening against Byzantine Exploitation in $q$-Fair Federated Learning
arXiv:2610.20264v1 Announce Type: cross Abstract: Client-level fair federated learning, in the form of $q$-fair federated learning ($q$-FFL), agnostic federated learning, or tilted empirical risk minimization, has emerge…
- arXiv cs.AIresearchCSWBench: Evaluating Context-Sensitive Code-Switching in Creative Writing
arXiv:2610.20272v1 Announce Type: cross Abstract: Large language models (LLMs) have demonstrated multilingual capabilities across diverse language generation and understanding tasks. Yet effective code-switching requires…
- arXiv cs.AIresearchBeyond Static Cues: Unified Dynamic Detection of MLLM Backdoors via Dual-Flow Trajectory Analysis
arXiv:2610.20319v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) are increasingly deployed in safety-critical domains, yet their susceptibility to backdoor attacks remains a formidable challenge…
- arXiv cs.AIresearchRead, Parse, Describe: Unified Document Parsing with Visual Element Description Generation
arXiv:2610.20335v1 Announce Type: cross Abstract: Document understanding systems are typically evaluated as separate components via layout analysis, optical text recognition, reading-order prediction, and image captionin…
- arXiv cs.AIagentsDoc2Harness: Benchmarking LLMs on Long-Horizon Harness Engineering of Coding Agents
arXiv:2610.20344v1 Announce Type: cross Abstract: Agents driven by large language models are increasingly deployed on long-horizon tasks, where they use tools and modify real workspaces. How well they perform depends as…
- arXiv cs.AIresearchNon-asymptotic Convergence of Average-reward Q-learning with Options
arXiv:2610.20347v1 Announce Type: cross Abstract: We study the finite-sample convergence of Q-learning algorithms for the average-reward formulation of Reinforcement Learning (RL) with options. While options allow tempor…
- arXiv cs.AIresearchHuman Demonstrations as Post-Training Data for Any Robot Body: Retargeting the Object, Not the Hand
arXiv:2610.20351v1 Announce Type: cross Abstract: Human demonstrations are today used largely for pretraining vision-language-action models, with little ability to use the same data for target-embodiment or task-oriented…
- arXiv cs.AIresearchCortex: Consolidating Context into Parametric Memory
arXiv:2610.20399v1 Announce Type: cross Abstract: Context can improve large language model performance, but these gains may be lost after context removal. Inspired by biological systems consolidation, we formulate Memory…
- arXiv cs.AIresearchWhen Does Momentum Help in Reinforcement Learning with Verifiable Rewards?
arXiv:2610.20413v1 Announce Type: cross Abstract: Momentum is a classical tool for accelerating convex optimization and a standard component of modern large language model training. Yet recent empirical work has question…
- arXiv cs.AIresearchOne Size Does Not Fit All: Curvature-Aware Adaptive Perturbation for SAM
arXiv:2610.20415v1 Announce Type: cross Abstract: Poor generalisation in deep neural networks is frequently associated with convergence to sharp minima. Sharpness-Aware Minimisation (SAM) improves generalisation of deep…