AI news archive.
19,861 articles, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchSame Probe, Different Numbers: Are Activation Probes Robust to Inference-Time Numerical Non-Determinism?
arXiv:2609.31796v1 Announce Type: cross Abstract: Activation probes are increasingly used to monitor LLMs in deployment. A probe is typically trained under one inference configuration, then used under whatever batch size…
- arXiv cs.AIresearchVideo-to-Music Generation for Gameplay Videos
arXiv:2609.31810v2 Announce Type: replace-cross Abstract: Video-to-music models have advanced considerably in the last few years, particularly in film and music video applications. In this paper, we investigate this prob…
- arXiv cs.AIresearchA Surgical Foundation Model Reveals Task-Dependent Label Efficiency
arXiv:2609.31821v1 Announce Type: cross Abstract: Developing label-efficient models is a central challenge in surgical AI due to the high cost and scarcity of expert annotation. While self-supervised foundation models ad…
- arXiv cs.AIresearchSynDORBench: Evaluating LVLM Perceptual Robustness Under Physically Constrained Visibility Conditions
arXiv:2609.31823v1 Announce Type: cross Abstract: Large vision-language models (LVLMs) have demonstrated remarkable performance on multimodal reasoning benchmarks, yet their perceptual reliability under physically constr…
- arXiv cs.AIresearchPHIRL: Aligning Learned Rewards with Task Progress for Inverse Reinforcement Learning
arXiv:2609.31855v1 Announce Type: cross Abstract: Human demonstrations provide dense policy-level information but sometimes lack local precision. Human feedback presents accurate local critiques, but offers sparse evalua…
- arXiv cs.AIresearchAirLog: Store-Level Indoor Life Logging Made Easy
arXiv:2609.31864v1 Announce Type: cross Abstract: This paper presents AirLog, a smartphone-based life journaling system that automatically reconstructs users' store visits in shopping malls and summarizes them into human…
- arXiv cs.AIresearchDeep Reinforcement Learning for Equity Trading: Benchmarking Actor-Critic Methods with Forward Retraining
arXiv:2609.31870v1 Announce Type: cross Abstract: Consistently profitable trading is difficult because equity markets are noisy, non-stationary, and only partially predictable from historical data. We benchmark five deep…
- arXiv cs.AIagentsCueKFS: Agentic Cue-Driven Keyframe Selection for Long Video Understanding
arXiv:2609.31873v1 Announce Type: cross Abstract: Keyframe selection (KFS) has long produced compact video summaries for browsing and retrieval, and representative frames for thumbnails. More recently, when conditioned o…
- arXiv cs.AIresearchA Large-Scale Benchmark and Risk Assessment of Traffic Analysis Attacks on Cloud LLM Services
arXiv:2609.31877v2 Announce Type: replace-cross Abstract: Cloud-based Large language model (LLM) services create a network-level traffic side channel that can expose model, prompt, and task behavior despite encryption. F…
- arXiv cs.AIresearchDOHF: Online Diffusion Fine-tuning with Doob's $h$-transform Guidance
arXiv:2609.31882v2 Announce Type: replace-cross Abstract: Reward-based diffusion fine-tuning faces practical challenges when desirable outcomes are rare or conditioning corrections are costly to estimate. In this work, w…
- arXiv cs.AIresearchFARE: Deep Reinforcement Learning For Fair Exposure Constrained Uncertainty Aware Financial Content Personalization
arXiv:2609.31890v1 Announce Type: cross Abstract: Content personalization systems in financial services must ensure fair exposure across diverse offerings-a requirement driven by contractual obligations and the need to p…
- arXiv cs.AIresearchNVAlign: Direct-Gradient Optimization for Non-Verbal Control in Continuous Autoregressive Flow Matching Text-to-Speech
arXiv:2609.31892v1 Announce Type: cross Abstract: While modern text-to-speech (TTS) systems generate highly natural speech and support inline non-verbal vocalization (NVV) tags, accurate control over these events remains…
- arXiv cs.AIagentsCyberWorld: World Models for Sample-Efficient Autonomous Cyber Defense
arXiv:2609.31893v1 Announce Type: cross Abstract: Deep reinforcement learning has become a prominent approach to autonomous cyber defense. Existing methods are predominantly model-free and consequently require extensive…
- arXiv cs.AIresearchAuditing Quality Filters for Long-Tail Human Data Curation
arXiv:2609.31896v1 Announce Type: cross Abstract: Robots on construction sites must detect workers who are kneeling or bending, which we call low poses. These workers can be lost from training datasets during automatic l…
- arXiv cs.AIresearchUnderstanding the Synergy between SFT, RLVR, and OPD in LLM Post-Training
arXiv:2609.31900v1 Announce Type: cross Abstract: Modern LLM post-training composes supervised fine-tuning (SFT), reinforcement learning with verifiable rewards (RLVR), and on-policy distillation (OPD) into multi-stage p…
- arXiv cs.AIresearchROTE: Benchmarking Neural Memorization on Complexity-Controlled Symbolic Sequences
arXiv:2609.31918v1 Announce Type: cross Abstract: We introduce ROTE (RollOut Testing of Exact memorization), a benchmarking protocol for evaluating symbolic memorization of neural architectures. We study memorization and…
- arXiv cs.AIagentsOverview of the TREC 2025 Million Large Language Models track
arXiv:2609.31921v1 Announce Type: cross Abstract: Agentic AI envisions ecosystems of intelligent agents collaboratively solving complex tasks with minimal human intervention. In such ecosystems, each agent possesses spec…
- arXiv cs.AIagentsVibe Analysis: Exploring LLM Adoption by Data Visualization Practitioners
arXiv:2609.31922v1 Announce Type: cross Abstract: Large language models (LLMs) are enticing in their promise to support data visualization (Vis) through faster and simpler workflows for data prep, analysis, and visualiza…
- arXiv cs.AIagentsVReAct: Verification as an Architectural Layer for LLM Agents: A V-Model Design, and a Pilot Study of Its Deterministic Core
arXiv:2609.31937v3 Announce Type: replace-cross Abstract: Large language model (LLM) agents built on the ReAct pattern concentrate four responsibilities in one model: selecting a strategy, choosing each action, formattin…
- arXiv cs.AIagentsCaptchaArena: A Large-Scale, Fine-Grained Dataset for Training Computer-Use Agents on Interactive CAPTCHAs
arXiv:2609.31957v1 Announce Type: cross Abstract: Interactive CAPTCHAs remain challenging for computer-use agents, while existing datasets face trade-offs among type coverage, interaction fidelity, and trajectory supervi…
- arXiv cs.AIresearchSNIP++: Fine-Grained Symbolic-Numerical Alignment for Symbolic Regression
arXiv:2609.31965v1 Announce Type: cross Abstract: Mathematical expressions and the numerical behavior they produce are two views of the same underlying function, and connecting them is central to scientific discovery. Sy…
- arXiv cs.AIresearchExtraction of clinical findings from mammography and breast ultrasound reports: a comparison between specialists and Artificial Intelligence
arXiv:2609.31974v1 Announce Type: cross Abstract: Breast cancer is the leading cause of cancer-related death among women in Brazil, and the time between the request and the release of mammography reports directly influen…
- arXiv cs.AIresearchCommunication between Frozen Large Language Models via Prompt Optimization in a Referential Game
arXiv:2609.31989v1 Announce Type: cross Abstract: We study communication between two frozen large language models from different providers, with different tokenizers, accessed through their API endpoints. The two play a…
- arXiv cs.AIresearchVC Dimension and Expressivity of Real-Valued Transformers
arXiv:2609.31999v1 Announce Type: cross Abstract: Whereas previous results on abilities and limitations of transformers have restricted the definition of transformers in various ways, here we study softmax-attention, mul…
- arXiv cs.AIresearchIs invariance all you need for algorithmic fairness? Removing demographic information can create new bias
arXiv:2609.32004v1 Announce Type: cross Abstract: Encoded demographic information in internal model representations is a commonly assumed risk factor for algorithmic bias, with demographic representation invariance often…
- arXiv cs.AIresearchTriO: Tri-Modal Unsupervised Occupancy World Model for Anything Perception
arXiv:2609.32013v2 Announce Type: replace-cross Abstract: We present TriO, a multi-modal unsupervised world model that predicts 4D occupancy, obstacle segmentation, flow and LiDAR. In contrast to prior work, TriO utilize…
- arXiv cs.AIresearchVoiceNet: Fine-Grained Voice Understanding Beyond Emotion at Scale
arXiv:2609.32016v1 Announce Type: cross Abstract: Expressive speech synthesis has outpaced expressive speech perception: systems now render fine-grained vocal performances that no public benchmark can score. Most benchma…
- arXiv cs.AIagentsSilentCall: Hidden Tool-Call Backdoors in Open-Weight Agents, and How to Catch Them
arXiv:2609.32021v1 Announce Type: cross Abstract: Open-weight tool-calling agents are adopted on evidence of merit, usually benchmark scores and a record of reliable use. We show that a model publisher can train an agent…
- arXiv cs.AIresearchDepth Any Seen: Which Surfaces and How Far?
arXiv:2609.32027v1 Announce Type: cross Abstract: When several surfaces are visible along a ray, recovering visible 3D structure from one image requires jointly estimating their presence and metric depth. Depth Any Seen…
- arXiv cs.AIresearchAmnesia by Design, Memory By Necessity: Persistent State for Document Intelligence
arXiv:2609.32041v1 Announce Type: cross Abstract: Modern Document AI reads contracts, extracts fields, reasons over tables, and grounds answers to page regions, then forgets everything. Processing an amendment the next d…
- arXiv cs.AIagentsInteractive Distributionally Robust Multi-Agent Learning with General Function Approximation
arXiv:2609.32048v1 Announce Type: cross Abstract: Model misspecification poses a fundamental challenge in multi-agent reinforcement learning, where transition uncertainty can be amplified by strategic interactions among…
- arXiv cs.AIresearchTracing Decoder Artifacts for Compact Synthetic Speech Screening
arXiv:2609.32050v1 Announce Type: cross Abstract: Recent advances in speech synthesis and voice cloning have increased the need for reliable synthetic-speech detection, yet high-accuracy detectors increasingly rely on la…
- arXiv cs.AIresearchReFM: Semantic-Aware Refinement Flow Model for Motion Retargeting
arXiv:2609.32068v1 Announce Type: cross Abstract: Motion retargeting transfers motion across characters with different skeletal structures while preserving semantic intent and physical plausibility. Despite recent progre…
- arXiv cs.AIresearchWhen Should a Human Take Back Control? Optimal Delegation under Turbulent AI Risk
arXiv:2609.32083v1 Announce Type: cross Abstract: Deploying AI systems requires deciding when to delegate tasks and when humans should intervene to monitor and mitigate risk induced by AI operations. These decisions are…
- arXiv cs.AIagentsOn Evaluating and Improving Conversational Agents in Production
arXiv:2609.32092v1 Announce Type: cross Abstract: We present a framework for evaluating and improving a large-scale, multi-agent shopping assistant in production, and report lessons from its use. Offline evaluation of su…
- arXiv cs.AIresearchEmergent One-Third Scaling Law as Attention Tries to Concentrate
arXiv:2609.32100v1 Announce Type: cross Abstract: The neural scaling law relating longer training to better performance through a power law is central to today's large language models (LLMs), yet its origin remains debat…
- arXiv cs.AIresearchEmpowering Hybrid Attention Models on NPUs
arXiv:2609.32114v1 Announce Type: cross Abstract: Hybrid attention models have emerged as a crucial architecture for Large Language Models (LLMs) (e.g., the Qwen3.5 and Kimi series). Their memory and computational effici…
- arXiv cs.AIresearchREALM: Regime-Switching, Explainable, and Activation-Induced Linear Models
arXiv:2609.32141v1 Announce Type: cross Abstract: Deep ReLU networks are piecewise-affine mappings that partition the input space into cells, each characterized by a distinct activation pattern. This structure motivates…
- arXiv cs.AIresearchSpectral Reversal: Counteracting Singular Value Bias for Graph Prompting
arXiv:2609.32143v1 Announce Type: cross Abstract: Pre-training Graph Neural Networks (GNNs) via self-supervised learning has become a dominant paradigm, yet efficiently adapting frozen encoders remains a challenge. Graph…
- arXiv cs.AIresearchTyped Decision Models: An Early Evidence Audit and Evaluation Checklist
arXiv:2609.32160v1 Announce Type: cross Abstract: Typed decision models (TDMs) return probability distributions over caller-defined options without generating text. TypeSafe released Jev, a commercial typed decision mode…
- arXiv cs.AIresearchUncertainty-Aware Selection of Online Algorithms with Simulator Ensembles
arXiv:2609.32170v1 Announce Type: cross Abstract: The performance of online reinforcement learning depends critically on design choices, especially those that affect exploration. These choices are often selected by fitti…
- arXiv cs.AIagentsOneFixer: High-Quality and Consistent One-Step Autoregressive 3DGS Refinement for Driving Scenes
arXiv:2609.32175v1 Announce Type: cross Abstract: Autoregressive video diffusion is a promising render-time fixer for 3D Gaussian Splatting (3DGS) in autonomous-driving simulation, but deployment demands high visual qual…
- arXiv cs.AIresearchKeyRec: Bounded Visual Memory for Streaming and Long-Video Understanding
arXiv:2609.32182v1 Announce Type: cross Abstract: Vision-language models are increasingly used to understand long videos and continuous streams. However, dense visual tokens accumulate with video duration, making long-co…
- arXiv cs.AIresearchPresence Is Not Faithfulness: Figurative Vehicle Intrusion in Text-to-Image Generation
arXiv:2609.32188v1 Announce Type: cross Abstract: Text-to-image (TTI) models increasingly generate high-quality images from natural-language prompts, yet figurative language exposes a failure: a vehicle that should guide…
- arXiv cs.AIresearchEvaluating Single and Multi-Omics Based Explainable Artificial Intelligence (MOXAI) for Molecular Subclass Classification of Adult-Type Diffuse Gliomas
arXiv:2609.32190v1 Announce Type: cross Abstract: DNA methylation (DNAM) profiling has emerged as a powerful diagnostic tool for classifying brain and solid tumors. However, existing computational models typically analyz…
- arXiv cs.AIagentsAgents as Software: A Programming Languages Agenda for Agent Reliability
arXiv:2609.32198v1 Announce Type: cross Abstract: AI agents increasingly resemble software systems: they call tools, remember facts, follow policies, delegate work, and take actions with real consequences. % Yet the ``pr…
- arXiv cs.AIresearchRank Confidence Sequences:Anytime-valid Leaderboards
arXiv:2609.32211v1 Announce Type: cross Abstract: Leaderboards rank models by their average scores on benchmark items, and they are consulted repeatedly while the evaluation is still running. Existing confidence interval…
- arXiv cs.AIresearchDP-Rec: Towards Dynamic Patching for Efficient Long-Sequence Recommendation
arXiv:2609.32215v1 Announce Type: cross Abstract: Transformers have redefined sequential recommendation by effectively modeling dynamic user behaviors and long-range dependencies. However, they remain inherently ineffici…
- arXiv cs.AIagentsLaMET-Agent: An Agent Framework for Large-Momentum Effective Theory Analysis
arXiv:2609.32225v1 Announce Type: cross Abstract: Large-momentum effective theory (LaMET) provides a first-principles framework for computing the $x$ dependence of light-cone parton distributions from lattice QCD. Over t…
- arXiv cs.AIagentsToward Agentic Optical Networks: A Vision of LLM Agent-Driven Autonomous Lifecycle Management
arXiv:2609.32226v1 Announce Type: cross Abstract: As optical networks continue to expand in scale, complexity, and service diversity, the implementation of automation has become essential for ensuring agility, efficiency…