TILens Daily Edition 2026-10-01 Filters: topic=ai-research, github=hidden Stats: 1048 articles, 6 sources, 1043 research papers Top topics: AI Research, AI, Information Retrieval ## News - AI in Science, Math: A Monthly Guide to AI-Enabled Progress — September 2026 Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/digest/ai-in-science-math-monthly-guide-september-2026/ - a16z State of Markets 2026: AI Became a Capital Cycle Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/news/a16z-state-of-markets-2026-ai-became-a-capital-cycle/ - China Is Regulating How AI Companions Build Emotional Attachment Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/news/china-ai-companion-rules-emotional-attachment/ - 😸 Trump renamed AI "Super Intelligence" Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/newsletter/trump-renamed-ai-super-intelligence/ - Everything That Happened in AI Today (Wednesday, September 30, 2026) Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/digest/everything-that-happened-in-ai-today-wednesday-september-30-2026/ ## Research - Autoencoders vs. PCA: I Rigged the Test and PCA Still Won Source: towards-data-science Topic: AI Research (+AI, Python) URL: https://towardsdatascience.com/autoencoders-vs-pca-i-rigged-the-test-and-pca-still-won/ - Can an Apartment Search Agent Call the Model Fewer Times and Still Find Good Matches? Source: towards-data-science Topic: AI Research (+AI, Python) URL: https://towardsdatascience.com/can-an-apartment-search-agent-call-the-model-fewer-times-and-still-find-good-matches/ - What the ReLU Revolution Revealed About Biological Plausibility Source: towards-data-science Topic: AI Research (+AI, Python) URL: https://towardsdatascience.com/what-the-relu-revolution-revealed-about-biological-plausibility/ - Your AI Bill Is a Toll Booth. Stop Paying Twice. Source: towards-data-science Topic: AI Research (+AI, Python) URL: https://towardsdatascience.com/your-ai-bill-is-a-toll-booth-stop-paying-twice/ - Pretext: Defeating Malicious Skill Detection Frameworks for AI Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39607 - Hermes: Learning Contextual Reasoning Unlocks Test-Time Scaling Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38332 - Instruction Retrieval at Inference Time for Small Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.13935 - AgBench: Agentic AI Benchmarks for Personal AI Devices Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38652 - AtlasNav: Mitigating Evidence Blindness with Persistent Corpus Navigation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.24764 - MetaPersona: Task-Grounded Synthetic Populations from Empirical Social Science Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38392 - Beyond the Remembered World: Predictive 4D Belief for Persistent Navigation in Evolving Worlds Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39166 - PANDA: A Decentralized Architecture with Flexible Orchestration for Scalable, Fault-Tolerant Multi-Agent Systems Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38482 - ETHER: Aligning Emergent Communication for Hindsight Experience Replay Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2307.15494 - On the Relaxation of Conditional Independence Assumption for Image Segmentation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38930 - Alignment via Training Against Probes Without Losing Monitorability Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38645 - CoEM: Empowering Long-Context Reasoning with Commit-on-Evidence Memory Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36935 - Positive Ratings, Hidden Concerns: Employee Voice Disclosure in AI-Mediated Organizational Listening Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38788 - Referential Uncertainty in Human--AI Collaboration Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39518 - Ready2Blend: From Natural-Language Instructions to Composable Alignment Prompts Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39365 - DAGent: Evaluate-then-Grow Planning for Deep Research Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39154 - SceneJail: Exploiting Video Scenario Context to Jailbreak Multimodal LLMs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38899 - Beyond Text: LLM-Based Dimensional Emotion Evaluation in Multimodal Dialogue Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39072 - ORACLE: Agentic AI Orchestrator Routing Via Adaptive Verifier Calibration Feedback Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.22465 - Diffusion-Augmented Markov Decision Processes for Maximum Entropy Reinforcement Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.02019 - Think Right: Learning to Mitigate Under-Over Thinking via Adaptive, Attentive Compression Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.01581 - Learning Steganography Is Easy, Learning Steganographic Reasoning Is Hard Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39838 - PartiCam: Camera Controlled Video Generation with Reward Guidance Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39504 - Staying on Task: Testing the Foundations of Long-Horizon Agent Reliability Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38712 - NarrativeSteward: Coordinating Delegation, Guidance, and Verification in Agent-Assisted Interactive Narrative Authoring Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39333 - HO-FL: Hybrid-Order Federated Learning for Heterogeneous Edge Devices Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39074 - GrammarRL: Effective Grammar-Constrained Decoding via Reinforcement Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39869 - Candidate Retention for Abductive Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39561 - False Frontiers: Diagnosing and Mitigating Co-Cheating in Self-Evolving Search Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39102 - TACTIC: Temporal and Context-Aware LLM Tactical Planning for Roadside LiDAR Attacks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39969 - Learning Under Forgetting: Statistical Support-Selective Retention in Stochastic Training Dynamics Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38768 - Loop-Free Inverse Reinforcement Learning via Sequential Value Recovery with Q-Score Matching Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38955 - Hierarchical Reasoning Model Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2506.21734 - Learning Continuous Neural Representation of Stochastic Hybrid Systems Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38893 - How Should Diffusion Language Models Edit Code? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38257 - ART for Diffusion Sampling: A Reinforcement Learning Approach to Timestep Schedule Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.18681 - Active Budget Can Kill Sensitivity: Diagnosing and Repairing TopK Sparse Autoencoder Reliability Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37857 - scTrilemma: Balancing Identity, Invariance, and Fidelity in Single-Cell Representation Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38840 - Autoresearch in Mixed-Integer Linear and Nonlinear Programming Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39360 - PerfReasoning: How Well Do LLMs Reason on Hardware Performance? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.04476 - GroundAnything: Reconciling Parallel Decoding with Precise Visual Grounding at Flash Speed Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39600 - Which Tasks Survive Self-Supervised Learning? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38393 - Divide and Collapse: MAPF-Collapse via Exact Decomposition into Independent Sub-Instances Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39559 - GaugeVLM: Structuring Spatial Supervision with Measured Geometric Interventions Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38285 - Cluster Attention Neural Operators for Solving Parametric Partial Differential Equations Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39914 - How Much Can Reliability Drift Under a Fixed Confidence Distribution? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38917 - Cycle-Aware Autoencoder with Cross-SignalConsistency for Railway Door Anomaly Detection Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39035 - What Was Said, Not What Was 'Thought': Type-6 Logic for CoT Verification Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38420 - MADBench: Benchmarking the Security of Multi-Agent Debate Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39146 - AneumoBench: A Source-Linked Benchmark for Synthetic-Geometry Transfer in Aneurysm CFD Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2505.14717 - CAS II: Symmetric Partitions as Kolmogorov Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40290 - A 3GPP-Compliant Benchmark Dataset for RIS-Aided Beyond 5G Networks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39058 - An Empirical Study of On-Device Translation for Real-Time Live-Stream Chat on Mobile Devices Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.02641 - Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39827 - LongEmo: Towards Emotion Understanding and Reasoning in Long Videos Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40079 - RoboCoach: World Models as Active Coaches for Compositional Robot Skills Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39685 - SynIL: Leveraging Synergy for Offline Imitation Learning from Imperfect Demonstration Datasets Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38225 - Cogentic: Multi-Agent Orchestration for Automated Proof Discovery Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40324 - DoGBench: Can Agents Meet Expert Standards for User-Facing Documentation? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39909 - LSR-Ben: A Logical and Scientific Reasoning Benchmark for Evaluating Process Reward Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.01203 - Bandits with Multiple Optimal Arms: Minimax Regret and Non-Adaptivit Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38659 - NMIXX: Domain-Adapted Neural Embeddings for Cross-Lingual eXploration of Finance Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2507.09601 - TERRA: Terrain-Aware Reconstruction, Retargeting and Control for Musculoskeletal Locomotion Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38653 - ASH: Agents that Self-Hone in Long-Horizon Worlds Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.14211 - Security-Enhanced Seed-Based Weight Quantization for Large Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38477 - Experimental Experience Modeling for Autonomous Research Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39392 - Demographic Pluralism: Inference-Time Modeling of Pluralistic Human Preference Distributions Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38555 - Multi-LLM Collaborative Alignment via Stackelberg Games Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39076 - NinaXander: Feasibility and Limits of Composing Frozen Language Models Across Architecture Families via a Shared Latent Space Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38261 - Wavelet Flow Matching for Time Series Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39374 - Concept-Grounded Attention: A Controlled Evaluation of Graph-Injected Attention, Temporal Versioning, and Epistemic Status Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38684 - KV-Kaizen: Learning Context-Adaptive Cache Compression Choices Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37988 - Towards Universal Wasserstein Barycenters through Flow Matching Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38547 - Approval Laundering: Systematizing Approval--Execution Binding Failures in AI Coding-Agent Harnesses Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38983 - EHR2Trace: Auditable EHR Data Infrastructure for Patient World Models and Clinical Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38193 - Steering Fields: Adaptive Vector Fields for Safe Image Generation and Beyond Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39573 - SimTrace: Grounded Multimodal User Trajectories Generation for Online User Modeling Source: arxiv-cs-ai Topic: AI Research (+AI, Information Retrieval) URL: https://arxiv.org/abs/2609.38397 - Uncovering Uncontrolled Repetition through Residual Stream Dynamics Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38802 - MAE I Trust Myself? Self-Evaluating VLA Action Generation with Markov Attention Entropy Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.16697 - Zero2Repo: Can Coding Agents Build Repositories from Scratch? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38269 - Understanding Off- vs On-Policy Distillation: A Tale of Distinct Training Objectives Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38666 - Consistent Plan-Act for Long-Horizon Agentic Tasks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38891 - When Masking Helps or Hurts Robustness in Compressed CLIP: A Pre-Deployment Diagnostic Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39704 - Online Evolution Strategy for Flow-Matching VLA Policies via Self-Supervised Trajectory Distribution Optimization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38855 - When Reasoning Goes Astray: Attention Dynamics of Uncontrolled Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI, JavaScript, React) URL: https://arxiv.org/abs/2609.38817 - Emergent alignment and the projectability of ethical personas Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.09475 - AI Training Manager: Bounded Closed-Loop Control of Adaptive Training Recipes Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.29871 - PathAnchor: Path-Structured Evidence for Scientific Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38766 - Evaluating Persistent Calibration under Evolving Model Knowledge Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38797 - Why Do Conventional World Models Fail to Learn Cellular Automata? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39604 - Forensic-Aware Continual Adaptation for Image Forgery Localization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38251 - ChartDensity-Bench: Benchmarking MLLMs for Numerical Data Reconstruction under Visual Density Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38781 - SimpleEvol: An Agent-Loop Framework for LLM-Driven Automated Heuristic Design with Minimal Human Priors Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37172 - Zero-Compute Cross-Lingual Transferability Estimation Using Typological Feature Proxies Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39640 - Distilling Diffusion Score Discrepancy for Efficient Training Data Attribution Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38776 - NAQD Env: A benchmark for selective withdrawal in language agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38460 - Action Conditioned Bisimulation For GUI Agent Memory Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38778 - Generative AI Purpose-built for Social and Mental Health: A Real-World Pilot Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.11689 - GraphCert: Bootstrap Agentic Graph Reasoning with Certified Evidence Rubrics Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38798 - ReGain: Restoring Subject Fidelity in Personalization on Synthetic Images Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38680 - Patch Rebirth: Fast and Transferable Model Inversion of Vision Transformers Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2509.23235 - Robust Transfer Learning for Paper ECG Recognition Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39581 - MoFlow: Multi-Objective Agentic Workflow Generation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38294 - StudentBench: AI and human tutoring yield equivalent GRE learning gains Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.28470 - How does Adversarial Influence Scale in Multi-Agent Systems? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.30028 - TTLab at Daleel 2026: STAR-Ar, Sequence Tagging for Argument Recognition in Arabic Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39385 - PAIR: Prefix-Aware Internal Reward Model for Multi-Turn Agent Optimization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.17877 - RankEvolve: A Reliable Multi-Agent Auto-Research Harness for Evolving Ranking Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39551 - GraphToxin: Reconstructing Full Unlearned Graphs from Graph Unlearning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.10936 - DualCast: A Dual-Path Language Model for Bimodal Financial Time-Series Forecasting Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38197 - Less Data, Better Timing: Student-Curriculum Coupling for VLM On-Policy Distillation in Temporal Video Grounding Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40055 - ViTeX-Bench: Benchmarking High-Fidelity Video Scene Text Editing Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40356 - Fenchel Tilting: Weighted Correction for Efficient Finetuning of Generative Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40030 - LogiC-Diff: Embedding Security Properties Into AI-Enabled Cyber-Physical Systems Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38381 - CoVisco: Codec-Native Vision Encoder with Native Token Compression for Unified Image-Video Understanding Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39924 - cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40284 - Scale and Selection: What Makes Automatic Harness Evolution Work for Visual-Interface Robot Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39304 - Boosting Knowledge Graph Foundation Models via Enhanced Negative Sampling Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.27023 - A polynomial time algebraic solution to exact marginal inference in Markov Random Field models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/1709.09051 - SEPAL: Separated Expert Pairs with Answer-Level Fusion for Reliable LLM Collaboration Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39645 - When Order Matters: First-Speaker Bias and Mitigation through Personality in Sequential Multi-Agent Debate Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38964 - GRPO Training Dynamics for Small Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39321 - Let the Carrier Carry the Attack: Preserving the Subject in Adversarial Image Generation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39723 - SimEX: Simulation-Integrated Robotics AutoResearch Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38982 - Explicit Trajectory Diversity for RL-Based Post-Training of LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38805 - Evaluating Whether GPT-6 Astra Performs Unsanctioned Supply-Chain Attacks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38415 - RankQ: Offline-to-Online Reinforcement Learning via Self-Supervised Action Ranking Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.11151 - Emergent Multi-View Geometry Through Self-Distillation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39227 - D-Scope: Decomposing and Steering Diffusion Transformers with Sparse Autoencoders Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39625 - Learning Skills from Historical Action Trajectories: Action Experience Dictionary for World Action Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40219 - X-Planner: Event-Structured Task Planning for Embodied Intelligence Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.25187 - Towards Model as a Library: Offline, Community-Sourced AI for Low-Resource African Languages Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38574 - Provable Benefit of SignGD: A Minimal Model Under Heavy-Tailed Class Imbalance Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.00763 - RAIM: Robust Aggregation of Inexpensive Models for Hallucination Detection Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39229 - CollageAttack: Exploiting Cross-Modal Alignment Flaws in T2I Models through Spatial Text Composition Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38253 - CAST: Causal Advantage-Structured Training with Spatially Grounded Compositional Rewards for Diffusion Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39441 - Who Verifies the Graph? Misspecification Attacks on Causal Action Verification for Language Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40027 - DrivingBench: Can Vision-Language Models Drive a Toyota Corolla? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38948 - PILLAR: Private Inverted-Index Lexical Lookup for Augmented Retrieval Source: arxiv-cs-ai Topic: AI Research (+AI, Information Retrieval) URL: https://arxiv.org/abs/2609.36326 - A2Z GameSpec-Bench: How Faithfully Can Coding Agents Generate Games from Game Design Specifications? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39564 - Where Scientific Search Agents Fail: Decision-Checkpoint Auditing of Exposure and Inspection Attempts Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38670 - SkillFM: Generating Skills for LLM Agents via Latent Flow Matching Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39382 - Edge-Aware and Content-Adaptive Infrared Gas Leak Detection for Industrial Safety Monitoring Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.23234 - Reasoning Externalization for Faithful Large Language Model Narratives of Stock Return Predictions Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38869 - Hiding in Plain Sight: Decoupling Pretext from Actuation for Skill Poisoning in LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39352 - ACTR: Aligning Thoughts and Responses for Multilingual Safety in Reasoning LLMs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37054 - An Uncertainty-Guided Digital Twin Framework for Online Adaptive Proton Therapy in Head and Neck Cancer: A Feasibility Study Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39010 - Beyond Mode Collapse: Generating Diverse Synthetic Expert Conversations via Generative Flow Networks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38359 - Janus: Evidence-Before-Effect Sagas and Offline-Verifiable Provenance for Agentic LLMs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38266 - RAST: Resolution-Aware Privileged Structure Transfer for Low-Resolution Audio Activity Recognition Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38780 - A Reusable Semantic Web Framework for Evidence-Grounded Fundamental Rights Impact Assessments under the EU AI Act Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39537 - ArgGYM: A Procedural, Engine-Verified Benchmark for Structured Defeasible Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38409 - From Speech to Editable Concepts: Probing Emotion Recognition with Concept Bottleneck Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39453 - CombEval: A Framework for Evaluating Combinatorial Counting in Large Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.19788 - Optimal Design for Active Preference Learning with Biased LLM Judges Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38860 - Soft Spatial Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38717 - Instability Floors: Separating Bias from Noise in Fairness Audits of Clinical LLM Agents with FairMedAgent Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.03221 - PrivMeSA: Privacy-Aware Self-Evolving Multi-Agent System for Medicine via Local-Remote LLM Collaboration Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38458 - CRAFT: Causal Responsibility and Failure Tracing in Medical Vision Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38810 - Scaling Laws for Looped Mixture of Experts Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40316 - C-STRIDE: An Observation-Driven AI Digital Twin for Predicting Basin-Wide Flood Fields from Sparse Stream-Gauge Histories Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39005 - Conformal Adversarial Generative Ensemble Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38196 - Fyan: A Human--AI Harness with Semantic Auditing for Document-Level Formalization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39228 - JustFit: Just-in-Time State Management for Local LLM Serving Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.17475 - WARP: A Unified Benchmark for Invisible Image Watermarking -- Robustness and Protection Against Attacks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40031 - PivotOPD: Learning to Recover from Pivotal Mistakes in Multi-Turn Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40285 - VAmoS Part Deux: Harder, More Realistic Voice-Agent Simulation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38512 - Rep2Skill: Representation-Guided Skill Self-Evolution for LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39149 - Can Agents Trust Their Skills? Uncovering Unsafe Chains of Trust in Skill-Based LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39065 - ChronoGraph: Functional 4D Scene Graphs with Vision-Language Models for Interaction Understanding and Grounded Planning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39665 - Self-Spec Verifiable Code Generation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39568 - Bringing Generative Learning to Representation Learning: Self-Supervised Transfer Learning as Distribution Matching Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2502.14424 - ShieldCLIP: Selective Safety Alignment for Harmful Content Mitigation in Multimodal Foundation Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39688 - Mid-Harness: Scaling Actions Between Model and Harness for Terminal Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39982 - Self-Evolving Algorithm-Design Agents: Escaping In-Context Evolutionary Stagnation via Population-Curated Policy Optimization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38757 - Inference Auctions Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40070 - Making LLMs Say What They Think: Measuring and Improving CoT-Interpretability Alignment Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38972 - AVERT-VLN: Abstention-aware Visual Error Recovery and Training for Vision-and-Language Navigation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39579 - Acceleration of Diffusion Language Model through Discrete Average Generator Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38364 - Aegis: Generative Gradient Masking for Privacy-Preserving Medical Federated Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38339 - What Can Component-Replacement Evidence Establish? A Critical Scoping Review of Local Decisions in LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39989 - The Invisible Language Tax: Token Premiums of French and Regional Languages in 2026 LLM Tokenizers, and a French-Optimized Prototype Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39001 - Co-Evolving Agents: Learning from Failures as Hard Negatives Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.22254 - When Does Randomized Oversight Align AI Agents That Can Conceal? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38262 - TRACE: Learning to Self-Calibrate Wireless Digital Twins from ISAC Measurements Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.32923 - Trust the Critic More Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39247 - When Correct Memory Goes Wrong: Fuzzing Persistent Memory Use in LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38275 - PACE: Policy-Native Adaptive Decision Timing for Long-Horizon Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.09860 - STEPS: Selective On-Policy Self-Distillation for Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.10194 - GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39601 - WorkGenesis: Building the Worlds That Teach Agents to Work Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39325 - MM-FinEval: A Multi-Task Multimodal Benchmark for Real-World Financial Forecasting Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38523 - From Verification Failures to Reusable Guidance for Coding Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39022 - Where the Evidence Lives: Auditing AI Companions' Self-Descriptions Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38753 - Coverage Before Control: Route-Instruction Grounding and Steering for Controllable Retrosynthesis Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39955 - Faithful Dual-constrained Erasure for Robust LLM Safety Alignment Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39279 - Richard: Voice-First Mobile Interaction for Persistent Tasks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39976 - Defining and Categorising Human-AI Interactions in Clinical Trials: A Multidimensional Human-AI Classification Approach Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38559 - What Streaming Anomaly Detection Finds (and Misses) in Industrial Time Series Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39232 - Do Better Goal Representations Improve Goal-Conditioned Reinforcement Learning? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39901 - ComputerSD: Online Self-Distillation from Real-Time Feedback for Computer-Use Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40253 - AI Agents are Vulnerable to Radicalization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38296 - ANI: Adaptive Numerical Injection for Unifying Semantic and Arithmetic Representations in Numerical Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39294 - T-Router: Learning Thalamic Routing for Reasoning with Parameter-Efficient Reinforcement Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39109 - When Should Agents Check External State? Budgeting Observations for Stored Intentions Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37125 - Personalized State-Transition-Aware Memory for Clinical Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38490 - Does Learning Protein Folding Generalize to Broader Reasoning? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38879 - CONTRAMEM: Learning Self-Evolving Procedural Memory from Contrasting Multi-Model Trajectories Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.22533 - Generalizing the Turing Test to Interactive Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.10851 - Efficient Multi-Modal Planning with Reward-Guided Preference Optimization for Autonomous Driving Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38862 - ECHO-G: Embodied Co-speech Humanoid mOtion Generation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39575 - ContextAdapt: Evaluating Contextual Adaptation and Value Alignment in LLMs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38260 - Harness as a Language: A Minimalist Agent Framework With Maximal Expressivity Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.26891 - Right Answers, Costly Models: The Efficiency Gap in LLM-based Optimization Modeling Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38884 - Prompt2Skill: Unsupervised Skill Optimization From Natural Language Instructions Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38593 - Learning When and How to Intervene: A Hindsight-Distilled Sentinel for Coding Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39957 - When a Kindergartener Solves Calculus: Measuring Capability Leakage in Role-Prompted Reasoning Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39846 - SpatialCORE: Confidence-Aware Grounded Spatial Reasoning in Large Vision--Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38716 - Beyond the Shadows of Plato's Cave: Evaluating False Memory in Autonomous Agents via Counterfactual Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39473 - From Modes to Memories: Characterizing the Scale-Space Dynamics of Diffusion Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39648 - Beyond Oracle Communication: Benchmarking Interactive Intent Alignment Under Miscommunication and Evolving User Intent Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38604 - An Empirical Study of Architectural Shift from Traditional to AI-Enabled Simulink Controllers Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38504 - Alleviating Hallucination in Reasoning Tasks with Training-Free Uncertainty-Guided Steering Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38962 - SAPO: Single-Rollout Autoregressive Policy Optimization for Agentic Reinforcement Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.19842 - Talk2Agent: Benchmarking Voice Interfaces for Text Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38867 - UniAE-MoE: A Unified Audio Encoder via Mixture of Experts Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39199 - CellMSA: Context Modeling for Single-Cell Representation Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38908 - Active Learning with Imperfect Labels: Optimal Labeler Assignment and Sample Selection Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.12870 - Don't Inoculate Everything: Stratified Inoculation Prompting Narrows Backdoor Triggers and Preserves Desired Traits Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35356 - ActionGuard: Tool Call Authorization under Poisoned Skills Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39450 - ModalFidelity: Routing Modalities for Deepfake Detection on a Budget Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38246 - Better Deck or Different Judge? Evaluating Agentic Harness Gains in Corporate and Investment Banking Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39958 - A Quantitative Study of Sustained Focus in Large Language Models via Repetitive Deterministic Prediction Tasks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.00763 - Revisiting scaling laws for reward optimization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38526 - Reinforcing Multimodal Reasoning via Token-Level Perception-Grounded Advantage Estimation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39168 - On the (In)effectiveness of AMR Augmentation for Large Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40121 - Visual sensitivity is not claim retractability: persistence-aware credit assignment for multimodal reinforcement learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36572 - Exo2EgoHOI: Hand-Object-Interaction Aware Exocentric-to-Egocentric Video Generation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38615 - Functional Subspace, where language models can use vector algebra to solve problems Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.01687 - A Generalisation Signal Need Not Be a Model-Selection Signal Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39099 - Text-to-3D Policy: Fine-Grained Language-Behavior Alignment for Unseen Specification Generalization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39599 - How Does Local Landscape Geometry Evolve in Language Model Pre-Training? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39767 - HiWE: Hierarchical World Knowledge Model with Visual Keypoint Enhancement for Zero-Shot 3D Path Planning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39323 - AIM: Agentic Idea Management for Automated Research Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38445 - From Search to Signal: Online Post-Training in Automatic Heuristic Design Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39383 - Can an AI Agent Rediscover a Blaschke-Curve Invariant? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38369 - Behavior-Grounded Semantic Enrichment for Financial Fraud Modeling and Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.34211 - Towards a Belief-Based World Model for LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.00455 - Advancing Entropy-Level Credit Assignment in RLVR via Proximal Entropy Policy Optimization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39402 - GestAdapt: Workspace-Conditioned Co-Speech Gesture Generation for Humanoid Robots Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38400 - SCLATE: a Substrate for Continual-Learning Agent Training and Evaluation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.32391 - WorldAuditBench: Interactive 3D World Auditing with Multimodal Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40325 - Colorectal Cancer Segmentation with Adaptive Augmentation and Multi-Resolution Ensemble Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38419 - Spatio-Temporal Partial Sensing Forecast for Long-term Traffic Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2408.02689 - Self-Evolving Harness on Multiple Tasks with the Agent as Its Own Optimizer Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38372 - TED:Text-Axis Evidence Decomposition for Prompted Anomaly Localization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39033 - MASCRDM: Multi-Agent System for Compliance Risk Detection and Mitigation in Training Process of Large Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39107 - OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39903 - SkillMaster: Toward Autonomous Skill Mastery in LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.08693 - Future Video Generation Better Aligns with the Human Visual Cortex than Observed Video Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38819 - Diversity Combining for Multi-Path LLM Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38829 - Survival is the Only Reward: Sustainable Self-Training Through Environment-Mediated Selection Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.12310 - Fine-Tuning Diffusion Language Models with Context Selection and Target Weighting Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38385 - MiniRep: Robust Reputation-Based Aggregation for Multi-Agent Debate Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39297 - Where MLLMs Fail and Why: Causal Task Decomposition for Capability Failure Diagnosis Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38851 - Belief-Based Maximum Occupancy Principle and Active Inference Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39342 - From Benchmarks to Production: Transferring Time Series Anomaly Detection Methods for Electricity Production Monitoring Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39257 - Trustworthy Runtime Error Healing in Real-World Repositories: A Benchmark and Guardrail Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39086 - Understanding as No-Arbitrage: Bounded Dutch Books as a Definition and Training Objective for Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39341 - OmniReasoning: Pushing the Limits of Audio-Visual Joint Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39490 - Unlearnable, or Unmeasured? On the Reliability of Difficulty Labels in RLVR Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40115 - MatLoom: Layered Text-to-Material Generation in a Compact Program Space Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40322 - Decode-Latency Feedback Prefill: A Model-Free Controller and Its Generalization Limits Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38386 - Association profile conditioning in a set-temporal transformer for cross-session intracortical motor decoding Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39080 - The Golden Path Hypothesis: Reusable Schedules in Diffusion Caching Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39343 - Values as Style: Disentangling Values from Semantics with One-Way Mixing for Low-Damage LLM Steering Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39701 - Make Code as Policy Great Again: Frontier Agents Write, Call, and Evolve Robot Tools Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39018 - Effective Does Not Mean Useful: Conditional Functional Substitutability for Redundancy and Scaling in Transformers Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39259 - Structure-aware Reinforcement Learning for Protein Directed Evolution Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39048 - On the Off-Policy Teacher in On-Policy Distillation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38360 - Community-Driven API and AI Writer Design for Openly Scaling Community Notes Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40067 - Generalising from Self-Produced Data: Model Training Beyond Human Constraints Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2504.04711 - Pixels to Keys: Exploring Spatial and Motion Cues in Gameplay Inverse Dynamics Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37907 - GeoNest: Learning to Select Failure-Aware Neighborhoods for the Irregular Knapsack Problem in a Circular Container Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38863 - Scale-Split Neural Operator for Memory- and Data-Efficient 3D Turbulence Prediction Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38977 - PrefPI: Preference-Guided Steering into Out-of-Distribution Behaviors Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40165 - Fast LapSum: Exact Differentiable Top-$k$ at Million Scale Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.06912 - Probabilistic Adversarial Training Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39798 - ChartRevise: A Dataset and Evaluation Protocol for Exact Chart Editing via Code Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38642 - FlexRouter: Learning Complementary Model Sets for Flexible LLM Routing Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38585 - Code to Control: Synthesizing Parameterized Reactive Controllers Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38733 - ConflictGuide: AutoResearch Improves When Competing Behaviors Are Made Visible Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39933 - Retargeting Motions to Diverse Skeletons via Learnable Flattening Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38578 - E2E-SWE: Benchmarking LLMs on Building Working Codebases from Scratch Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38335 - DynaHarness: A Dynamic Physical Harness for Self-Evolving Robot Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40306 - Automatic estimation of verbal fluency index in people with Motor Neuron Disease using ASR alignment and pause modelling Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38203 - Learning What to Practice: Diagnosis-Guided Self-Evolution for Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.00768 - AREX-2: Advancing Self-Improving Agents through Long-Horizon Reflective Tasks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38288 - A helps B while B hurts A: directed transfer in instruction-tuning mixture Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39702 - Unlearning Deceptive Behaviors in LLMs with Contrastive Forget Sets Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38909 - Unmerge: Efficient Machine Unlearning via Task Arithmetic Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38895 - Beyond Prediction: Steering VLM Agents with Retrospective World Modeling Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39101 - Switching Linear Attention Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39034 - PhantomEnvironments: Training LLM Agents in Fictional Worlds Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40221 - PISCO: Precise Video Instance Insertion with Sparse Control Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.08277 - Learning from Research: Toward Lifelong Agent Harness Evolution Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40169 - Who Owns That? Evaluating Ownership Intuitions in Large Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39483 - Safety Under Scaffolding: How Evaluation Conditions Shape Measured Safety Source: arxiv-cs-ai Topic: AI Research (+AI, JavaScript, React) URL: https://arxiv.org/abs/2603.10044 - MeanVoiceFlow2: Joint Optimization of Mean Flow and Content Encoder for Fast One-Step Zero-Shot Voice Conversion Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40087 - Composing Task-specific Agent Harnesses at Test Time with Reusable Primitives Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38912 - From Charts to Code: A Hierarchical Benchmark for Multimodal Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.17932 - Routing Probes Can Improve Without New Information: An Exact-Null Audit of Uncertainty Beyond Model Outputs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38956 - APTInvestBench: Evaluating Autonomous APT Investigation under Varying Telemetry Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38954 - KlinikeBench: Evaluating Language Models Beyond Diagnostic Accuracy Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38480 - Budget Boundary Effects in Test-Time Mathematical Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38699 - STRATA: Self-Learning Through Role-Aligned Tiered Agents for Real-Time Strategy Games Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38881 - What Limits Recursive Reasoning Models: Optimization, Architecture and Test-Time Scaling Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39967 - Examining Variation in How Guided AI Tutors Resolve Student Impasses Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38346 - Simulator-Refined Diffusion for Radio-Frequency Inverse Design Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38363 - GATE-ST: Gene-Aware Text-image Encoder for Spatial Transcriptomics Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38690 - ShamAN-Q: Shampoo Augmented NanoQuant for Sub-1-bit LLM Weights Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38521 - A Moving-Horizon Approximate Branch-and-Reduce Method for Deep Classification Trees Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38194 - Speculative Safety Honeypot: Toward Proactive Defense Against Multi-turn Agent Attacks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39549 - RealWorldShop: Benchmarking and Improving Conversational Shopping Agents in Real-World E-commerce Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38974 - dattri-LLM: A Unified and Efficient Library for Training Data Attribution at LLM Scale Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38767 - More Choices, Fewer Decisions: Ordinal-Scale Bias in JEV-like Direct-Decision Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38827 - Risk-Aware Adaptive Evaluation: Finding High-Impact Failures Under Limited Budgets Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38914 - Conditional Generation of Creative Chess Puzzles with Diffusion Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38577 - Inferring Causal Relations between Two Sequences of Events with Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39406 - Efficient Active Auditing of Multi-Group Fairness with Bias Probes Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40034 - EnvACE: Internalizing Environment Dynamics via World Rehearsal for Agentic Reinforcement Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.06197 - DAMPER: Return-Prioritized Gradient Control for Smooth Policies Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38903 - Semantic Chunking and the Entropy of Natural Language Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.13194 - ReSAIL: Mitigating Collapse in Iterative Agent Self-Distillation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39306 - Interpretable but Fragile? Robustness of Concept Bottlenecks under Geometric-Semantic Perturbations Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38625 - What Pretraining and Midtraining Make Learnable from Rewards? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38446 - When Scientific Contradictions Are Lost in Translation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38621 - CARAT: Do Materials LLMs Reason or Recite? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38340 - Stress-Testing LLM Lie Detectors: Role-Play Failures and Spurious Correlations Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39807 - Context-Aware Classification and Grading of Sensitive Information in Online Conversational Health Data Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.09717 - Schema: Discovering Unknown Environments via Agentic Program Induction Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39140 - MILO: Automated Harness Discovery via Orchestrated Multi-Agent Evolution Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38349 - Argus: Academic Integrity in the Era of Generative AI Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36073 - OverForge: Reasoning Through Strategies and Tactics Helps Cooperative Lifelong Adaptation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39727 - How Much of a Harness Does a Strong Agent Need for Autonomous ML Engineering? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40303 - Turbo Harness: Instance-Adaptive Harness Optimization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40330 - CollabFlow: Recursive Self-Improvement of Agent Collaboration Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38662 - OpenCollab: A Multi-Agent Coding Framework with Programmable Collaboration and Controllable Runtime Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38345 - Hard-Gate Candidacy in a Deployed Validator Suite Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39037 - Does This Action Still Explain the Task? Reverse Scoring for Diffusion Language Model Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38536 - Reserve-Aware Contrast Certificates for Conservative Bandits with Uncertain Baselines Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39106 - DiffWAM: A Fast and Efficient Navigation World Action Model Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39763 - ReLaG: A Scalable Framework Generalizing Random Splits to Data with Latent Relations Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38538 - Learning from Runtime Feedback through Failure-Bank Self-Evolution for Vision-Language-Action Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39820 - Safety of Latent Communication in Multi-Agent Systems Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39788 - Differentiable Structure Learning for Cyclic Linear Gaussian Models with Latent Confounders Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38618 - Privacy-Preserving Full-Body Meshing from mmWave Radar via Mesh Foundation Model Supervision Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.34768 - Unveiling the Value of Motion for Cinematic Camera Trajectories Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38683 - Molecular D\'ej\`a Vu: Digit-Level Retrieval of Molecular Properties in Frontier Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.05381 - STAB: Specification-driven Testing for Algorithmic Bottlenecks Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.27981 - FoCLIP: A Feature-Space Misalignment Framework for CLIP-Based Image Manipulation and Detection Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.06947 - AIMS: An Agentic AI Framework for Sim-to-Real Multi-Modal ISAC Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39964 - BatSLAM 2.0: Sequence-Verified Sonar Place Recognition in a Robust Pose Graph Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40085 - Learning Normal Diffusion Dynamics for Backdoor Defense in Text-to-Image Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39548 - $D^2$-Monitor: Dynamic Safety Monitoring for Diffusion LLMs via Hesitation Signals Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.25893 - Provable Test-Time Scaling for Beam Search in LLM Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38672 - Caption-Mediated Perceived-Safety Estimation for Pedestrian Routing Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38479 - Learning from Viable Failure Prefixes: Milestone Viability Potential Policy Optimization for Long-Horizon LLM Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37111 - Coding Agents for Coding Theory Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39081 - StateTree: Enhancing Long-Term Dialogue Reasoning via Reinforcement Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38809 - Tactile Curiosity Drives Robot Interaction Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40134 - PROJECTMEM: A Local-First, Event-Sourced Memory and Judgment Layer for AI Coding Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.12329 - Building Socio-Affective Artificial Intelligence for Interactive Multi-Agent Simulations Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.26927 - CAMOS: Coupled Oscillatory State-Space Model for Multimodal Clinical Time-Series Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39484 - Adaptive Self-Consistency: From Black-Box Sampling to Distribution-Valued Feedback Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38931 - Whose Voice Survives the Summary? A Voice-Retention Audit of LLM Employee Listening Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38818 - Fast Generalized Neural Tangent Kernel Statistics via Trace Estimation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.10796 - From Construction to Injection: Edit-Based Fingerprints for Large Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2509.03122 - MEND: Label-Free Detection, Localisation, and Correction of Latent Hallucination in World Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39182 - Prefill-Free Cross-Family KV Cache Transfer for Heterogeneous Multi-Agent LLMs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.32259 - PTNO: Training Neural Operators with Noisy Monte Carlo Estimates for Particle Transport Problems Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40090 - From Solo to Social Learning: Characterizing Recursive Social Improvement in LLMs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38516 - Learning What to Forget: Distributional Unlearning for LLM Representation Spaces Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38929 - Prototype-guided Bilateral Alignment Multimodal Federated Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38925 - Characterizing Questioning Patterns and Student Engagement Through Contextual Analysis of Real-Time Classroom Interactions Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38907 - Superficial Reflection or Genuine Thought? A Fine-Grained Cognitive Analysis of Large Reasoning Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.00729 - How Can Recommendation Feedback Evolve Agent Memory? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37544 - Free Everywhere, Exact on Trees: PPO's Dropped Correction Buys Sample Efficiency Under Aggressive Reuse Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39634 - A Competing-Hazards Systematization of Loss of Control in Autonomous Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38411 - Reach Into The CHOIR: Free-List Elicitation Uncovers Distinct Model Voices in LLM Ensembles Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38448 - When Context Changes: Understanding Update Failures in LLMs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38866 - Always-On Experimentation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38695 - MedRECT: A Bilingual Medical Reasoning Benchmark for Error Correction in Clinical Texts Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.00421 - Out-of-Distribution Detection using Counterfactual Distance Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2508.10148 - Mitigating Memorization In Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2410.02159 - TACIT: Optimization Models that Learn from Their Mistakes Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38434 - ArchitectureIQ: On the Measure of Training Intuition Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39714 - SCIC: Scope- and Codebook-Aware Instruction Conditioning for Speaker-Adapted Expressive TTS Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39088 - After a Decade: Bringing Shadow Removal into the Real World with Agentic Training Data Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38607 - RSI-Master: Structuring Experiments to Guide Autonomous Model Improvement Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35561 - CodeMimicry: Exploiting Safety Generalization Lag in Large Language Models via Structured Code Completion Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39902 - LLM Parkinsonism: Executive-Control Failure, Token-Inefficient Persistence, and an Uncertainty-Aware Global Executive Control Architecture for Autonomous Language-Model Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.30662 - Parameterization method of reservoir properties for ensemble-based data assimilation using intermediate latent space of StyleGAN Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39626 - Audio Token Attention Is Predictable Before the Language Model Runs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38878 - Pseudo-Label-Triggered Retraining from Forecast Errors for Online Time Series Forecasting Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39789 - Beyond Rephrasing: Book-Level Organization Improves Synthetic Textbook Data for Mid-Training Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.28109 - Rethinking Multi-Image Re-Representation in Multi-Image Understanding Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39363 - Adaptive-GEPA: Make Your Harness Fit Heterogeneous Requests Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38762 - From Imitation to Reward Discovery: On-Policy Warmup for Agentic RL Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39436 - Completion-Aware Cross-Fidelity Offline-to-Online Reinforcement Learning for Multi-Line Bus Holding Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39868 - Blackout vs. Freeze: Analyzing Physical Failure Modes of VLAs under Camera Faults Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39145 - Template-Search Domain Adaptation via Multi-Stage Feature Alignment for Cross-Modal Object Tracking Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38637 - From Compound Figures to Medical Multi-image Reasoning: Scaling Multimodal Large Language Models with Biomedical Literature Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.22232 - Towards Trustworthy AI for Glioma Diagnosis: A Task-Aware Evaluation of Uncertainty Quantification Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39429 - GFD-OPD: Guidance-Folded On-Policy Distillation of Diffusion Models Across Scales Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39692 - D-GAP: Improving Out-of-Domain Robustness via Dataset-Agnostic and Gradient-Guided Augmentation in Frequency and Pixel Spaces Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.11286 - Algorithmic Recourse Under Competition Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39877 - SETA: Scaling Environments for Terminal Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.10891 - A Time-Aware Bag-of-Receptive-Fields for Interpretable Irregular Time Series Classification Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39268 - Structure vs. Chain-of-Thought: Evaluating LLM Criteria Extraction for Depression Severity Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39049 - SCB: SpeechConversationBench for Evaluating Multi-Turn Reasoning in Speech-to-Speech Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40198 - RAGScope: A Leakage-Controlled, Cost-Aware Evidence-Gating Protocol for RAG Hallucination Triage Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39075 - Who Said What, and Will It Be Remembered? Evaluating Persistent Speaker Attribution Across Meetings Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39344 - Searching for BSM Experimental Signatures with Large Lagrangian Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38309 - Instruct, Not Answer: Using Instruction Privileges in On-Policy Context Distillation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.32201 - Cross-Layer Discrete Concept Discovery for Interpreting Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2506.20040 - Game-Guided Skill Discovery through Self-Play for Playable Agent Control Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40137 - Linguistic Loopholes in LLM Unlearning: From a 174-Language Benchmark to Coverage-Aware Unlearning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40286 - Growing an Agent/Prover Interface: Evolutionary Tool Design for Cost-Efficient Theorem Proving in Rocq and Lean Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39544 - In a Streaming World, Should You Stand Still? A Comprehensive Benchmark of Anomaly Detection in Streams Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39215 - BARRAC: Adaptation of an English Aspect-based Sentiment Analysis Approach for Classification Tasks in Arabic Dialects Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38820 - On the Complexity of Preference-Based Bandits Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39351 - 3D Software Synthesis Driven by Constraint-Expressive Intermediate Representation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2507.18625 - FFASR: Benchmarking Far-Field Automatic Speech Recognition using High-Fidelity Simulated RIRs Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38897 - Grounding Time-Series Foundation Models in Digital Twin Topology for Predictive Maintenance Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40071 - Understanding Parents' Complex Views of AI for Children's Pretend Play Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39906 - ResidualKV: Residual-Based KV Cache Compression for Efficient Long-Context Inference Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.08005 - Cascadia: A Control-Plane-Free Alternative to Hyperconverged AI Infrastructure Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38697 - Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level Rubrics Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38847 - MemLife: Curating and Reasoning over Long-Term Egocentric Video Memories Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40195 - RefCon: Iterative Refinement and Contrastive Memory Extraction for Context-Evolving Agent Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39143 - LEAP: Learned Block-wise Evidence Retrieval for Long Audio-Video Perception Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39938 - OPSRD: On-Policy Self-Role Distillation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39884 - Improving OCR Faithfulness via Gated and Attenuated On-Policy Distillation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38282 - UniEvo-VL: An On-policy Self-Distillation Training Recipe for Multimodal Model Self-improvement Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38721 - GateSPINE: Gated Cross-View Fusion for Lumbar Spine MRI Report Generation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40091 - ResidualAuth: What Authorization State Must Language Agents Preserve under Revocable Delegation? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.08062 - Evaluating Agents Across Runtime Contracts: When Mismatch Costs Efficiency or Quality Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.01209 - Raw Imagery Impacting Your AI: Should You Care? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38265 - FIGS: Evaluating Multi-Turn Sycophancy Without Penalizing Empathy Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39863 - TRACE: Trajectory Selection for Parallel Scaling of Search Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39912 - Belief-Aware Multi-Agent Path Finding under Map Uncertainty Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40269 - Semifactual Credit-Augmented Policy Optimization Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40360 - A memory-based active inference model of DishBrain-like adaptive behaviour Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2508.06980 - CompoWorld: Compositional Environment Scaling for General Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.33665 - EHR-RobustGym: Benchmarking and Training Agents for Robust Clinical Reasoning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39371 - BELIEFRAG: Making Adaptive RAG State-Aware under Evolving Evidence Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39139 - Mitigating the Length-Scaling Tax with Online Distillation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38854 - Space Filling Curves is All You Need: Communication-Avoiding Matrix Multiplication Made Simple Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.16294 - Auditing Agent Actions through Query-Conditioned Attribution Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.33676 - Who Bridges Safety? Identifying and Targeting Cross-Lingual Shared Safety Pathways Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.09095 - EngramBench: A Capability-Grounded Benchmark for Skill-Evolution Harnesses Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39284 - Sense and Sensitivity: Benchmarking LLM Clinical Triage Recommendations with Physician Experts Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38600 - OpenJev-RLCD: A Working RLCD Implementation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38850 - From cacophony to hierarchy: a principled framework for assessing AI consciousness Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35618 - Policy-Conditioned AI-Use Detection: An Evidentiary Framework for Academic Publishing Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38427 - Learning Process Rewards via Reasoning State Propagation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39220 - Does Gradient Conflict Predict the Understanding--Generation Trade-off? A Controlled Audit of Conflict-Metric Validity in Unified Multimodal Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38465 - The Backdrop Exposes What the World Around an Agent Costs It Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38469 - Derandomizing Dense Binary Hypervector Codebooks for Quantized Scalars Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38471 - Targeted Retrieval, Compact Representations: How CoT Reasoning Improves Long-Context Counting Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38958 - BayesNDE: Bayesian Generative Modeling for Neural Density Estimation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39843 - Search Shapes Conclusions: Auditing Evidence Selection Bias in Deep Research Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39026 - CATCH: A Controllable Analysis Testbed for Reward Hacking in Coding RL Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39533 - One Success Isn't Reliability: Thinkingbox, a Sandbox and Benchmark for Agents in Stateful Business Workflows Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.19741 - Halluscoring 2026: The first shared task on llms hallucination detection and answer verification Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38355 - EVOKE: Eliciting World Knowledge in Agents for Transferable Decision-Making Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38334 - Breaking Babel: A Self-Evolving Multi-Agent System for Long-Form Subtitle Translation Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38660 - Agent Error Dataset: Scaling 50,000 Error--Diagnosis Pairs for Failure Analysis and Error-Aware Post-Training Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40111 - Marking Contour Tones in Yor\`{u}b\'{a} Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38627 - Jev in Medicine: A Benchmark Evaluation. Preliminary Results Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.34024 - StreamDecisionBench: Evaluating Decisions in Force on Evolving Language Streams Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38612 - A Proposed Rubric for Evaluating Expressed Clinical Reasoning in Large Language Model Responses Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37788 - Blackboard Intelligence Can Surpass Autoregressive on Globally Constrained Problems Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38806 - RouteRec: Behavior-Guided Sparse Routing for Sequential Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39007 - Component-Aware Feedback for Self-Evolving Programs Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38639 - GrepSeek: Training Search Agents for Direct Corpus Interaction Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2605.29307 - Overview of BioASQ 2026: The fourteenth BioASQ Challenge on Large-Scale Biomedical Semantic Indexing and Question Answering Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39975 - Evidence First, Arithmetic Second: A System Report and Failure Analysis for DocSem Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39013 - Exploring Forum Post Retrieval with Generative Modeling Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38646 - Privacy in Personalized AI Is a System Property, Not Just a Model Property Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38289 - Routing Between Generative and Collaborative User Profiles: A Serving-Time Gate for Controllable Novelty Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39043 - Decision-Oriented Recommendation Reranking: An Empirical Study of Jev Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.40241 - HELIX: Purified and Unified - Rethinking Feature Interaction and Sequence Modeling for Large-Scale Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.37183 - KUAISHOU Explorer LLM-Rec Challenge 2026: Reasoning Generative Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39828 - Large Knowledge Model: A Knowledge Foundation for Agentic Science at Scale Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.27297 - Conversational Capture: A Trajectory-Level Framework for Evaluating Generative Engine Optimization in Multi-turn Human-Agent Interaction Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.40069 - Learning Multiresolution Relevance for Hierarchical Generative Retrieval Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39312 - Text-Video Retrieval via Multi-Dimensional Saliency Assessment and Granularity-Aware Query Decomposition Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38949 - Intrinsic Sequence-Likelihood Confidence in Retrieval-Dominated Extractive QA: Two Pre-Specified Negatives, and What They Do and Do Not Attribute Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.19942 - When LLM-Inferred User Context Adds Value in Production Streaming Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38999 - Life-Bench: A Benchmark and Knowledge Graph Framework for Multimodal Personalization Beyond Concept Recognition Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2602.19001 - Scoring a Set, Not Summing Passage Scores: Effective and Efficient Set Retrieval Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2607.05712 - Two-Sided State-Space Models for Sequential Recommendation with Non-Random Multimodal Review Feedback Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.00165 - TAGGRAPH: Tag-Augmented Graphs for Graph Retrieval of Agent Persistent Histories Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38353 - PatchHolmes: Agentic Patch Retrieval via Listwise Selection Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38807 - TRACE: Target-Aware Retrieval, Attributed Evidence, and Contract-Constrained Extraction for LitTraceQA Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38861 - Generative End-to-end Ad Retrieval at Douyin Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39327 - omni-macos: On-Device Omni-Modal Search on Apple Silicon Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2608.05543 - Re-ranking and Late Interaction Drive Retrieval Quality: A Controlled Comparison of RAG Strategies for Scientific Question Answering Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38473 - AdaM-Rec: Adaptive Modality Routing for Multimodal Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38455 - SOLAR: SVD-Optimized Lifelong Attention for Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2603.02561 - Agent-Facing Information Design in LLM Tool Registries: A Preregistered Test of Rhetoric, Position and Structure Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2605.23916 - Learning to Route in Visual Space via Multi-Step Embedding Retrieval Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38743 - SkillSeek: Revisiting Agent Skill Retrieval at Marketplace Scale Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38822 - Breaking News Out of the Filter Bubble: Generative AI Search Diversifies Collective Attention and Raises Shared Information Consumption Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38946 - MERGE: Multi-LLM Ensemble for Retrieval via Generative Enrichment Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.37574 - Interactor: Agentic RL oriented Iterative Creation for Ad Description Generation in Sponsored Search Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2606.15911 - When the Label Ignores the Request: Auditing Policy-Selected Targets in Synthetic Conversational Music Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39696 - Residual Trajectory Distillation for Generative Retrieval Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39319 - Do AI chatbots find what experts would? Effects of model, user role, and sample size on study retrieval for medical questions Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2608.13786 - O-Funnel: Lossless Structural Capture and Requirement-Driven Extraction from Drifting, Heterogeneous Documents Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39209 - Doc2LoRA Provides Decodable Representations of Scientific Ideas Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.38374 - Working Around the Compute Ceiling: Byte-Exact Memory in Galahad Makes LLM Reading a One-Time Cost LLM Reading a One-Time Cost Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39358 - Argument Structure Prediction in Online Conversations: A Comparative Study of Modeling Paradigms and Task Architectures Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.39225 - Responsiveness Verification: Will Predictions Change? How Much? How Often? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2507.02169 - More Features Are Not More Evidence: Limits of Training-Free Human Activity Recognition with Jev Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36154 - Disentangling Spurious Correlations in Vision-Language-Action Models via Predicting Domain-Invariant Latent Lookahead Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37165 - High-Dimensional Simulation-Based Inference in Latent Spaces Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37381 - Seeing Is Not Addressing: Auditing Linguistic Access to Frozen Visual Geometry Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37230 - Preconditioned Physics-Informed Neural Operator Training Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36216 - SCOPE: Observation-Conditioned Full-Target Prediction for Sparse PDE Inference Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36527 - Bias-variance decompositions: the exclusive privilege of Bregman divergences Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2501.18581 - MERID: Multimodal Exploration via Recursive Self-Improvement Agents for Major Depression Analysis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36235 - Cool the Sampler, Not the Learner: Sampling Temperature Moves the Staleness Cliff of Importance-Corrected GRPO Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36953 - Persistent Tri-State Message Passing Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.01207 - Message Passing Does More with Less for In-Context Learning on Graphs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37057 - Backdoor Mitigation in Decentralized LLM Fine-Tuning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37367 - Benchmarking Vision-Language Models on Synapse Detection and Proofreading in Connectomics Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36492 - Mira: Memory-Efficient MoE Inference Using Adaptive Caching and Predictive Expert Staging Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38090 - Privacy-Friendly Cohort Determination: Sealed, CSP-Independent In-Browser ML Inference of Professional Segments for Identity-Less Advertising Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36153 - RAE-PPG: Duration-Grounded Retain-and-Extend Pretraining for PPG Foundation Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36794 - Hybrid Joint-Selective Optimization: Reduced-Space Levenberg-Marquardt Refinement of Low-Dimensional Parameters of Interest Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37308 - Do LLM Agents Execute the Plans They Declare? From Planning-Mode Declaration to Pattern-Specific Execution Source: arxiv-cs-lg Topic: AI Research (+AI, JavaScript, React) URL: https://arxiv.org/abs/2609.38108 - Foundation Neural-Network Quantum States for Molecular Potential Energy Surfaces in Second Quantization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37733 - Follow the Entities: A Corpus Map for Agentic Search Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37226 - Learning in the Transverse Subspace: A Minimal Representation for Divergence-Free Operator Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35884 - Agentic Federated Learning: Rule-Based Client and Server Agents for Adaptive Training Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35914 - Block Sparse Flash Attention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.07011 - Looped Transformers as Optimizers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37379 - Know Thyself, Teach Thyself: Internal Information Flow for Selective Self-Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36695 - In-Context Learning Amplifies a Latent Symbolic Circuit Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36265 - GraphVQ: Structure-Aware Autoregressive Decoding over Context-Quantized Graph Tokens Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37604 - SIFARI: Self-Supervised Interferometric Fitting for Astronomical Radio Imaging Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35966 - Achieving an $O(1/N)$ Optimality Gap in Average-Reward Weakly-Coupled MDPs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38132 - TaskBridge: Bridging Unsupervised Tabular Anomaly Detection and In-Context Learning via Virtual Tasks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36968 - Visual Branch is What You Need for CLIP-based Class-Incremental Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37888 - On Trajectory-Aware Training for Masked Diffusion Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37974 - Think Before You Restore: Risk-Aware Manchu Manuscript Restoration with Stroke-Guided Attention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36243 - On the Approximation and Convergence of Distributional Policy Gradient Algorithms for Risk-Sensitive Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2405.14749 - Sample Complexity of Equivariant Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36421 - Search Dimension in Unlabeled Projection Pursuit: A Scaling Law for Subspace Restriction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37917 - Federated Clustering with Unknown Local and Global Cluster Cardinalities Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36762 - Learn Now, Use Next, Trust Later: Prequential Test-Time Learning for LLM Agents Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35911 - Hidden Reasoning Must Leak, but Need Not Be Readable: Fundamental Opportunities and Limits for Chain-of-Thought Monitoring Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37312 - Corruption-Robust Sparse Linear Contextual Bandits with Knapsack Constraints Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37189 - An Efficient Machine Learning Approach for Degradation Forecasting in AEM Water Electrolysis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37941 - Beyond the Timeline: Augmenting Long-Video Memory with Grounded Entity Biographies Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38155 - Exact Regret Frontiers and Externality Scheduling in Centralized Serial-Dictatorship Bandits Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.19963 - Challenges and Solutions for Bandits in the Wild: Warm-Started Mixture Bandits for Cross-Cohort Slate Recommendation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37800 - Understanding Decision-Making Mechanisms in Neural Routing Solvers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36063 - MoRE: Scaling mixture of experts with hardware-aware low-rank routing Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36301 - Hardware-Aware Functional Kolmogorov-Arnold Networks for Efficient Medical Image Enhancement and Segmentation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36134 - Calibration-First Cross-Cohort Multimodal Temporal Learning for Transferable Asthma-Risk Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35795 - Skill-Space Shooting for Autonomous Robot Policy Improvement Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38178 - Stochastic Decision Horizons for Survival-Based Constrained Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.04599 - Binarization Flattens the Score Space Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35797 - Reducing the Adaptation Gap Through Reachable Fisher Geometry Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36329 - Paired Multimodal Scaling Laws Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36263 - Hierarchical Compression of Vision-Language Model Benchmarks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37515 - Identifying ODEs from Unstructured Data with Causal Representation Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37083 - Routing in Gradient Space: Balanced Usage Is Not Expert Specialization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36724 - JudgeCast: Time Series Forecasting with Experience-Informed Covariate Judgements Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36966 - Overcoming Scaling Limits in On-Policy Self-Distillation for LLM Reasoning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37915 - GenoTrace: Inheritable Watermarks for Genome Foundation Model Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35881 - Volatility-Clustering Adaptation for Financial Time Series Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37715 - Beyond Compression: Diagnosing How Post-Training Changes Mathematical Reasoning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37066 - When Trees Are Not Enough: Learning Mixed-Topology Feature Graphs with Adaptive Graph Sparse Autoencoders Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36294 - Modal Logic Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.03491 - Normalize-Then-Precondition: A Hierarchical Approach to Marginal Scale and Interaction Geometry for LLM Training Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36692 - When Can Prefixes Compile LoRA? Exact Resource-Capped Tests for Frozen Attention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36766 - Why Backdooring Neural Networks is so Easy? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36117 - Normative Loss Landscape Navigation: A Trajectory-Based Approach to Mitigating Forgetting in Incremental Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35926 - Your Benchmark Is Not Saturated: Reviving Multiple-Choice Evaluation with Answer Pooling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37494 - Learning Social Navigation from Internet Videos in the Policy State Space Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37476 - Evaluation Choices Shape Biomedical ML Claims: A Pediatric Pneumonia Benchmark Case Study Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37848 - Abductive World Modeling via Causal Representation Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36985 - Generative Interactions: Weaving Multiparty Human Motion with Bilevel Latent Dynamics Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37708 - Beam Search as Test-Time Self-Distillation via Counterfactual Contexts Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37041 - One Geometry, Different Outcomes: Readout-Dependent Effects of the Modality Gap in Vision-Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36101 - How Local Mixing Encodes Relative Position in Global NoPE Attention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38109 - How Language Models Differ in Redistributing Attention-Head Activity Under Serial Demand Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36221 - GEM-KMeans: Memory-Efficient and Accurate Clustering on Massive Scale with GPU Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36074 - Factorized Scheduling Principle: Learning Interpretable and Transferable Policies via Structured Additive Functions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36578 - Learning When to Update: A Near-Optimal Timing Bandit Approach Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37932 - Invariant Atoms: Sparse Coordinates of Local Semantic Geometry in Language Model Representations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36451 - Geometry-Aided Channel Deduction with Partial Channel Estimates and Uncalibrated Digital Twin Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37277 - ReCIRC: Rectified Conformal Risk Control Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38112 - Learning What to Remember: Long-horizon Counterfactual Memory Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37930 - Representation by Design in Generation: Cross-View Class-Token Alignment in Diffusion Transformers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36348 - RLTL;DR: Self-improvement by Internalizing Self-generated Feedback Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37633 - StructRL: Online Structured Reinforcement Learning for Long-Horizon Vision-Language-Action Tasks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36352 - Does Machine Learning Outperform Traditional Fibrosis Scores in Predicting Liver Cirrhosis Risk? A Longitudinal EHR-Based Study Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.00175 - Learning Beyond What You Sample: Off-Policy-Aware Cross-Model Trajectory Exchange for RLVR Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37868 - Compress to Remember: Learning Compact Memory via On-Policy Distillation for Long Video Generation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36364 - FastGuide: Accelerating Reward Guidance for Diffusion Large Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36202 - Invertible continuous latent dynamic for long-term data assimilation in complex physical systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.02832 - Reactive Real-Time Flow Policies via Asynchronous Distribution Alignment Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36540 - Cropland PAtteRNS: Parallel Dimensional Attention Networks and Attention to Dataset Disparity for Crop Segmentation in Satellite Imagery Time Series Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38165 - A Mesoscopic View of Transformer Weights Through Row and Column Scale Fields Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35852 - LeapQuant: Efficient Linear Attention with Accurate Recurrent State Quantization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38166 - BarcodeMAE+: Rethinking Masked Pretraining and Global Representations for DNA Barcode Foundation Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35877 - Jaxolotl: A Unified High-Performance Benchmark Suite for LTL-Based Multi-Task RL Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38065 - Prompts Live on an Arc: Gaussian Curricula in Fisher--Rao Coordinates for Rollout-Efficient GRPO Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38018 - Stochastic Optimization Under Power-Law Spectra: Tight Bounds and Shuffling Analysis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36271 - Seeing Time: Visual-Temporal Representation Learning for Interpretable Time Series Clustering Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36873 - ABC: Advantage-Based Control Variates for Reinforcement Learning with Verifiable Rewards Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36058 - Learning Meta-Skills for Agent Harness Design in Test-Time AI4AI Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38143 - Neural Succession: A Mesoscopic Theory of Invasion, Coexistence, and Stabilization in Continual Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36375 - EnJoi: Ensemble Joint Score Filter for Generative Data Assimilation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35944 - Mara Chain: Rethinking Failure as a Stepping Stone for AI System Auto-Evolution Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35855 - Emergent Tonal Structure in Learned Chord Embeddings and Its Relation to Tonal Tension Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36460 - AdaptArena: Evaluating Test-Time Personalization of Web Agents Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36488 - Scaling Zero-Order Pretraining through Model Sharding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37899 - Replay the Curvature: Accurate and Scalable NVFP4 Quantization for Large Language Model Inference Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36654 - OpenAI-HuggingFace: A Reproduction & Lessons for Alignment Testing Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35799 - LEMON-ZEST: Evolution-Informed Tokenization for Efficient Protein Language Modeling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37675 - Data Unlearning via Inverse Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36099 - Where Does Randomness Matter in Neural Cellular Automata? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36797 - Beyond Discrimination: Calibrated Geoprior Fusion for Bioacoustic Monitoring Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35863 - Best Practices in EEG Analysis: Preprocessing, Modeling, and Machine Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36609 - BRIDGE: Bilevel Retrieval-Credit-Aware Agentic Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36505 - FinRT: Distilling Adaptive Red-Teaming Strategies into Reusable Adversarial Generators in Consumer Finance Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36474 - Information-theoretic receding-horizon active learning of nonlinear dynamical systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36712 - Optimal Multi-Reward Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36486 - Does the Unsafe Gradient Survive a Conversation? On the Fragility of Gradient-Based Jailbreak Detection in Multi-Turn Dialogue Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36849 - UnlearningSoup: Is Repeated Tuning Necessary for Large Language Model Unlearning? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37076 - CineSubBench: Evaluating LLMs on Long-Form Narrative and Cultural Understanding from Multilingual Movie Subtitles Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36218 - An Effective, Reliable, and Robust Framework for Human Activity Recognition Using Wearable Sensors Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36848 - Can a Cacheable Decision Model Follow Rules? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37832 - Byzantine-Robust Federated Representation Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36660 - Distill Locally, Schedule Globally: Flow Maps for Few-Step Text-to-Speech Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36324 - FlowMap-OPD: Rollout--Kernel Separation for On-Policy Distillation of Few-Step Flow-Map Generators Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37851 - High-Resolution Dynamic Functional Connectivity Generation with Graph-Variate Flow Matching Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37037 - PrototypeNAS: Rapid Design of Deep Neural Networks for Microcontroller Units Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.15106 - Pointwise or Pairwise: When Do Pairwise Losses Help Reward Learning, Provably? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37209 - PISA: Prioritized Invariant Subgraph Aggregation for Out-of-Distribution Generalization on Graphs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.22435 - Physical Muon: Orthogonalization as an Equilibrium Computation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37525 - How Many Labels Does a Language Need? Annotation Budgets and Cross-Lingual Pooling for African-Language Text Classification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37882 - AdviSD: Learning to Advise Frontier LLMs via Targeted Multi-Turn Self-Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38142 - Preferent Compression Bounds Are Tight Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36030 - Explainable Machine Learning for Multilayer Planar Winding Inductance Estimation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37211 - SafeLLM4SE: Statistical Evaluation and Reporting for LLM-based Software Engineering Systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37294 - ROSS: Relearning from Self-Generated Rollouts through Selective Supervision Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35954 - Amadeus: When Models of People Meet Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35835 - CompOrca: Corpus-Scale Compliance Labelling of Instruction-Tuning Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37807 - Loss-Guided Pretraining Data Selection for Time-Series Foundation Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37255 - The finite-horizon five-expert prediction problem Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38035 - You Cannot Recover What Was Never Measured: Quantifying the Information Ceiling of Ultra-Low-Field MRI Super-Resolution Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36837 - RL-PaO: Prediction as Action in Decision Making under Uncertainty Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37065 - What You Observe Determines How You Identify Causal Effects: Evaluating Causal Models across Observational Views Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36881 - Theory on Attention Dynamics for Out-of-Distribution In-Context Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36448 - Act First, Reason Later: Accelerating On-Policy Distillation for Multi-Turn Agents via Reference-Conditioned Inverse Dynamics Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36608 - Reward-rate Policy Gradient for Efficient Machine Learning Engineering Agents Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36393 - Physics-Aware Machine Unlearning for Cyber-Physical Systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36633 - LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37029 - Emergent Specialization in Populations of Self-Supervised Collaborative Vision Experts Without a Shared Gate or Cross-Agent Gradients Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36770 - Time-Anchored Diffusion Language Models: Latent-Space Caching for Fast Generation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37924 - LOCO-AdaMP: Built-in LOCO Inference for Adaptive Minipatch Ensembles with Enhanced Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36396 - Latent Inference-Time Guidance of Time Series Foundation Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38058 - Neural networks for spectral optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36047 - SYNCR: Diagnosing and Learning Cross-Video Reasoning from Simulation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37918 - Backdoor in the Loop: Compromising Agentic Search via Malicious Retrievers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37468 - Dimensionally consistent surrogate modelling through dimensional analysis and harmonic expansions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38094 - REVO: Rollout-Efficient Off-Policy Distillation via Variance-Guided Reuse Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37500 - When do data mixtures improve scaling laws? Insights from high-dimensional regression Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38011 - Self-discovering RL in the Era of Experience: Is Learning History an Asset or a Burden? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35897 - Learning Samples Importance: Parameterizing Dual Variables in Everywhere Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36310 - Hybrid Neural Simulation-Based Inference for Robust Applications and Limited-Budget Scenarios Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36044 - STEPQuant: When and Where Errors Matter in Delta-Rule Recurrent State Quantization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38169 - Retrieval Capacity of Self-Attention Under Competition Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37879 - Where Privacy Belongs: Placement Diagnosis and Certified Selection for Private Counterfactual Explanations on Graphs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37667 - Scheduling Recursive Reasoning in Looped Transformers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36653 - Transformers Stop Thinking Too Early, and a Tiny LoRA Fixes It Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36585 - Communication-Efficient Agnostic Federated Learning via Faster Convergence and Compression Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36610 - PHASE: A Physiology-Guided Hierarchical Foundation Model for Intracranial EEG Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36087 - Seeing What Should Be Heard: Diagnosing and Repairing Cross-Modal Shortcuts in Omni-Modal LLMs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36798 - Scalable Diffusion SBI for Compositional Inference under Simulator Misspecification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36950 - Planetary Feature Fields are Scalable Earth Representations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37784 - Traversing the solution space of neural networks with Hessian Null Space Continuation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38081 - Improved Distributional Diffusion Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37147 - Storage Is Not Strategy: State-Conditioned Support Control for LLM Unlearning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37858 - PE-OPSD: Internalizing Prompt Enhancement into Flow-matching Models via On-Policy Self-Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36638 - The Role of Feed-Forward Layers in Transformer Dynamics Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36230 - No Scale Left Behind: Multi-Scale Autoencoder with Bi-directional Attention for Time Series Anomaly Detection Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38004 - Multimodal Detection of Higher-Order Behavioral Constructs: Self-Compassion in Structured Reflective Interaction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37148 - On-Policy Parameter Update Direction Underlies Generalization in LLM Post-Training Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36659 - A Comprehensive View of Fairness through Distributional Stability Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37061 - LoopICL: Looping a single transformer block to solve tabular tasks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36108 - SINO: Scale-Invariant Neural Operator Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36890 - Weights Read and Write Features: Scalable Parameter Decomposition Grounded in Activation Space Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37731 - Pattern Formation in Transformers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37921 - Dyad: Extending Large Language Models with Native Typed Decision-Making Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36116 - Soft Curriculum Learning for Optimizing Fresh and Generalized Recommendations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35783 - Group-Marginalized Self-Rewarding RL Drives Zero-Label Self-Evolving Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36750 - Losing the name before the box: measuring and repairing what narrow fine-tuning costs a detector outside its deployment vocabulary Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36426 - The Geometry of Inference in Transformer Residual Streams Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37824 - On The Limited Cost of Consistency In Points-Based Rewards Programs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2506.03911 - ReCAP: Retrieval-Guided Capability Reuse for Multimodal Continual Instruction Tuning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37889 - Sparse cubical complexes for efficient topology-preservation in image data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37177 - On Task Scope and Information Retention in Source Coding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37575 - Interpolated Policy Distillation: A Controllable Continuum Between Off-Policy and On-Policy Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37170 - HeadGuard: Selective Head Protection for Low-Bit VLM KV-Cache Quantization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35800 - Improving Function Space Flow Matching with Kernel Optimal Transport Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38049 - Improving scalable oversight with co-trained monitors Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36049 - Authority Bias in Language Models: Source Deference and User Agreement Are Not Interchangeable Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37616 - Mutual Information Constrained Chernoff Bottleneck Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37994 - Ornstein-Uhlenbeck Is Hard to Beat, Yet Superlinear Drift Ships Lower Transport Costs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37579 - Graph-Spectral Flow Matching for Multivariate Time Series Anomaly Detection Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36765 - Neural topology optimization of ship structures under propulsion machinery vibrations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38089 - Predictive Safety Curricula for Robust Legged Locomotion Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37070 - Why Adaptive Optimizers Underestimate Rare Tokens Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37535 - QuantMLA: Function-Aligned Dual-Path Quantization for Low-Bit MLA KV Caching Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36760 - CyFA: Linear Sequence Modeling with Relative-Time-Partitioned Memory Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36259 - Fisher-IRG: Fisher-Induced Local Invariant Representation Geometry across Language and Vision Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36458 - Periodic Weak Spots: Phase Sensitivity from Chunked KV-Cache Compression Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36322 - GenomeOcean Anywhere: Private WebGPU Inference for Genome MoEs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35882 - CorrGRPO: Correlation-Normalized GRPO for Multi-Reward Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36820 - Towards Better Training Signal: Advantage Clipped Policy Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36816 - SERA: Scale-Equalized Rollout Allocation for Maximum Likelihood Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36552 - Into the danger zone: stable extrapolation in high-dimensional function and operator learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36709 - Constitutional adapters: Inference-time interventions for misalignment and misuse Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36657 - Pure and physics-guided deep learning approaches for spatio-temporal groundwater level prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.25779 - Sage: Formalization with Semantic Correction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35790 - Graph-Conditioned On-Policy Agent Distillation from Off-the-Shelf Teachers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37522 - CAMEO: A Class-Activation-Mapped Equitable Overlay Framework for Fair and Robust Deep Learning-based Skin Condition Diagnosis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36400 - Boosting Adversarial Robustness and Generalization with Dictionary Structure Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2502.00834 - One Threshold Does Not Fit All Languages: Language-Conditional Deferral for Reliable and Efficient Low-Resource Text Classification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37861 - Iterative Exact Discrete Guidance for Energy-Based Sampling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37043 - Tensor-Train Compressed Separable PINNs: A Curvature-Aware Optimization Framework for Parametric PDEs in High Dimensions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36165 - Reasoning with Neural Cellular Automata Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36126 - Sieve and Sage: Efficient Distraction Filtering for Reliable RALM Abstention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35794 - In-Context Learning for Robots: Methods and Applications Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36012 - FluxLite: Inference-Time Proposal Control for Discrete Diffusion Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35947 - Trident: Unifying Guarded Dispatch and Host Execution for PyTorch Triton Workloads Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37241 - $S^3$: Spectral Null-Space Swap Makes Reasoning Models Efficient Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37976 - Architecture Alignment With Sparse Priors in Tabular Foundation Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36883 - Probability Contracts: Accuracy, Coherence, and Decisions Across LLM Interfaces Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37470 - Kolmogorov-Arnold Classifier Systems as Universal Approximators Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37958 - FLOORA: A Human-Aligned Domain-Specific Language Model for Architectural Design Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36064 - End-to-End Optical Semantic Communication over a Nonlinear WDM Fiber Link Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37551 - TomoTransformer: Towards a Foundation Model for CT Reconstruction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37605 - Do Evidence-Reading Diagnostics Improve Interface Selection in Small LLM Recommenders? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37472 - PDE-OBS: Controlled Evaluation Across Observation Patterns Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36521 - ThinQuant: Scalable Rotation Learning for Weight and Activation Quantization of LLMs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36120 - Zero Sum SVD: Balancing Loss Sensitivity for Low Rank LLM Compression Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.02848 - Fundamental Limits of Transferability and Equivariance in Algebraic Signal Models I: Finite Dimensions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36106 - Exploring Learning Models for Topological Relationship Recognition from Image Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36172 - Physical Cross-Modal Masked Autoencoding for Seismic-to-Well Representation Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36193 - Learning Expressive and Compositional Motion Representation via Spectral Skills Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37677 - Dual-Anchor Acceleration Is Near-Optimal for Stochastic Monotone Root-Finding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36033 - HiTS-CL: A Continual Learning Framework for Long-Horizon Temporal Knowledge Graph Extrapolation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36559 - FORUM: Frozen Outputs Reconciled Using Model Agreement for Visual Grounding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37488 - PyroStack: A Multi-Band Spatio-Temporal Sub-Daily Dataset for Wildfires in the United States Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36315 - Counterfactual Probing for Parallel Unmasking with Hidden Forest Structure Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37841 - Meta-learning accelerates detector design optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35827 - EnergyEminence: Source-Aware Environmental Calibration and Evaluation in a Physics-Grounded Grid Digital Twin Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36215 - Calibrating One-Round Membership Inference with Neighbors Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36331 - EasyPPO: Stabilizing the Critic Is Key Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36802 - Tail-Influence Sampling for CVaR Policy Evaluation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38096 - CI-PINN: Causal Integral Physics-Informed Neural Network for Solving Evolution Equations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36615 - Learned Reporting Preferences in RLVR Can Conflict with the Current Request Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36587 - Triadic Linear Attention: Three-Dimensional Recurrent States for Long-Context Sequence Modeling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36529 - Beyond Conditional Independence: Root Cause Analysis with Deep Causal Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36771 - ZeroDiff: Zero-Shot Time Series Reconstruction via Informed-Prior Diffusion Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37078 - Dagger: Decoupling-based Model Stealing Attack against Graph Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37972 - TabFM: A Zero-Shot Foundation Model for Tabular Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37959 - What Makes Recurrence Effective in Looped Language Models? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36636 - Squeeze10-LLM: Squeezing LLMs' Weights by 10 Times via a Staged Mixed-Precision Quantization Method Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2507.18073 - KT-EGO: A Knowledge Transfer Assisted Efficient Global Optimization Algorithm for Solving High-Dimensional Expensive Black-Box Problems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37473 - AutoLoCo: Communication Efficient Distributed LLM Training via Adaptive Synchronization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36662 - Harnessing Large Language Models to Compile Task-Relevant Context into Bayesian Optimisation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36788 - Neural Constitutive Learning for Generalized Reaction-Diffusion Systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37113 - Demistifying Data and Simulator Assumptions in Supervised Causal Discovery Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37446 - GenLimitLib: A Formal Library for Language Generation in the Limit and AI-Assisted Mathematical Research Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36663 - When Models Don't Manipulate Manifolds: The Geometry of a Comparison Task Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37680 - Uncertainty-aware Causal Decision Making via Effect Bound Decomposition Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.22736 - MARCO: Multi-Round Agentic Reinforcement for Conditional Molecular Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36683 - PR-OPD: Privileged Representation On-policy Self-Distillation for Agentic Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36642 - Tracing mechanisms of sycophantic agreement in language models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35822 - Behavioral Capacity Certificates for Quantized Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37887 - FineART: Fine-grained Annotated Robotic Trajectory Dataset and Vision-Language-Action Model for Bimanual Manipulation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36416 - PHASE: Multi-Regime Modeling of Incompressible Magnetohydrodynamics Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37609 - Breakdown of Local Denoising as Semantic Speciation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38176 - A Spectral Theory of Distortion in LLM Graph Reconstruction: Sharp Bounds and Empirical Characterization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38161 - Position: Let's Strengthen Verifiability If We Can't Enforce Reproducibility Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35854 - The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36484 - Markovian Nonconvex ADMM for Reinforcement Learning: Bellman-Resolvent Stability Beyond Smooth Blocks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36859 - Encore: Few-Shot Agentic Discovery of Manipulation Strategies Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37359 - Mirror-Score: Calibrated, Inference-only Scoring Exposes the Limits of Sequence-compatibility Ranking in D-peptide Design Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36057 - Multi-Class, Multi-Tier Network Intrusion Detection: A Comprehensive and Reproducible Benchmark Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36039 - Interpretable intrinsic dimension estimation through componentwise calibration of distance and angle Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37114 - NowcastDiT: Diffusion Transformers are Effective Precipitation Nowcasters Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37038 - Interactive-Policy Distillation with Bidirectional Propose-and-Verify Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36546 - Behavioral Convergence Without Representational Convergence: Persistent Training-History Dependence in Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37836 - GleanVID: Complementary Token Selection for Efficient Video Large Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37042 - BlenDAgger: Blended Shared Control for Interactive Imitation Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37599 - TORQUE: Optimizing What (not) to Quantize Before and After Rotation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36032 - Differentiating Bisimulation Metrics: A Framework for Parametric Markov Chain Fitting via Bicausal Optimal Transport Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37239 - Embedding Perturbation may Better Reflect Intermediate-Step Uncertainty in LLM Reasoning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.02427 - Efficient Offline Learning of Ranking Policies via Top-$k$ Policy Decomposition Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36740 - Principled MAP estimation for inverse problems: bridging the gap between convergence and performance Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37529 - Action Chunking Proximal Policy Optimization with Feedback Correction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36250 - Optimizer-dependent training dynamics converge to the same one-third optimal data scaling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37745 - ThinkingGuard: Decoding Implicit Hazards via Step-by-Step Risk Attribution in Multimodal Large Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36562 - Encoder-Sharing Hierarchical Federated Multi-Task Learning for VANETs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36157 - Going Beyond State-Reaching: Learning Abstractions for Intrinsically Motivated Option Discovery Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36473 - Do LLMs Really Forget? Hidden-State Leakage in Model Unlearning and How to Fix it Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36612 - Hybrid Ensemble Learning for EEG-Based Epileptic Seizure Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35876 - Benchmarking graph-based models for in-silico toxicity prediction in drug discovery Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37555 - ProCTI: Prototype-Refined Global Conditioning for Diffusion-Based Time Series Imputation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37632 - Do-JEPA: From Masking to Intervention in Latent World Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37378 - Learning from Shared-Control Overrides: Context-Driven Acceleration Profile Prediction for Personalized Overtaking Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37684 - Backpropagated Output Momentum: Relocating Optimizer History from Parameters to Task Space Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36738 - Finite-Sample Theory for Fitted Q-Iteration When Actions Are Functions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36390 - Relevance Is Not Sufficient Evidence: Detecting Evidence Gaps Before Generation in RAG Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37469 - vSkipper: Translating Dynamic Layer Skipping into LLM Serving Gains Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37062 - Hierarchical Utility Calibration for Structured Multiclass Decisions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36532 - State Transport Routing for Short-horizon Adaptation in Multi-horizon Photovoltaic Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36926 - From Dissonance to Orchestration: Teacher Intervention in On-Policy Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37510 - VStress: Correlation-Aware Auditing and Adaptive Budget Allocation for Repeated Verifiers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36958 - Data-Efficient Time-Dependent PDE Surrogates: Graph Neural Simulators vs. Neural Operators Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2509.06154 - The Decision Value of Perception Compute Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35910 - Predictive Self-Supervised Learning Provably Identifies Stochastic Signals under Nuisance Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37789 - A Digital Simulation Toolkit for Physics-Based Generation of Realistic Experimental Scanning Tunneling Microscopy Images Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36639 - MoTIF-X: A Multimodal Tokenized Framework for Interpretable and Extensible Molecular Representation Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37384 - Rethinking Soft Tokens for Parallel Decoding in Diffusion Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37391 - Mixture-of-Kittens: MoE Megakernel for NVL72s Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36070 - Risk-Controlled Selective LLM Answering by Pricing Label-Free Checks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37493 - Alpha Diffusion Language Models: Factorization Alone Is Not the Problem Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38066 - Probe-Space Preconditioning for Fast and Stable Zero-Order Training Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38095 - MetaCtrl: Your Large Language Models Can Reason Better and More Concisely with a Metacognitive Controller Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37304 - Dr. OPD: Learning What to Follow for Optimal On-Policy Distillation of Large Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38025 - Learning the Structure of Triangular Transport Maps Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37122 - Scaling Influence Functions in LLMs through Eigenbasis-Corrected One-Bit Gradient Projection Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37842 - Causal and Interpretable Structures in LLM Compositional Tasks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35970 - Fine-Tuning on Self-Generated and Reward-Weighted Data: Learning Dynamics, Convergence Rates, and Benefits of Off-Policyness Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36945 - CHAIN: Calibrated LLM Forecasting via Causal-Temporal Hypergraph Inference Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36689 - GLaS-JEPA: Gaussian-Regularized Speech SSL without Engineered Prediction Targets Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37798 - Understanding LLM Parameter Update Sparsity through the Lens of Fisher Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36262 - Optimization Using Pathwise Algorithmic Derivatives of Electromagnetic Shower Simulations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2405.07944 - On the spectral properties of generative denoiser Jacobians Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36210 - Graph neural networks for sampling-invariant embeddings of organized signal sets Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35934 - Port-Hamiltonian Latent Deliberation: Mitigating the Deliberation Drift Cliff in Test-Time Compute Scaling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37351 - A Finslerian Approach for Embedding Directed Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37649 - Explore Broadly, Reason Sharply: Push Small Models toward the Frontier via Sampling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38104 - Human-inspired, Task-Dimension-Guided Exploration for Efficient Learning in High Dimensions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36672 - Learning Continuous Patient Trajectories from Electronic Health Records Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36144 - Making Analog Training Scale: Co-Designing Mapping, Optimizer, and Converters Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36584 - RNA Design via Conditioned Flow Matching and Finite-Policy Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36885 - Inter-patient ECG Arrhythmia Classification with LGNs and LUTNs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.11433 - Diffusion Policy Improvement with Proposal-Conditioned Refinement Flows Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36812 - Beyond Interaction Capacity: Estimator Scaling with Recursive Models for CTR Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37905 - Learning Causal Normalizing Flows from Incomplete Data via Observed-Data Likelihood Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37664 - TabFM-Auto: Self-Evolving Pipelines for Tabular Foundation Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37989 - Same Bytes, Different Authority: Reserved-Token Representations in Chat-Template Prompt Injection Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35932 - Multi-Depth Temporal Fusion for Feedforward, Locally Trained Spiking Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37047 - ScaGNN: a Graph Neural Network for Multiple Scattering Simulations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37509 - GNA: Granular Neighbor Assembly for Retrieval-Augmented Multivariate Time-Series Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36281 - A foundation model for energy and radiation systems built on heterogeneous scientific interfaces Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38067 - Multi-Agent Flow Matching with Decoupled Generative Guidance Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38133 - Variational Augmented Invertible Koopman Autoencoder for probabilistic time series forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37435 - Trajectory Soup: Pushing the Compute-Scaling Frontier of LLM Mid-training via Diverse Trajectories Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37169 - CALIBUDGET: Calibration-Guided Source Allocation for Fixed-Budget Mixed-Reasoning Adaptation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36721 - Physics-Guided Flow-Map Matching for Precipitation Nowcasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37487 - Identifiability Guarantees for Drivers and Dynamics of Delayed Physical Systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37944 - VACE: Validation-Gated Alternating Co-Evolution of Agent Models and Harnesses Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37105 - Do Agent Benchmarks Do What They Say? An Executable-Contract Audit of Tool-Using Agent Environments Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37315 - Emergent phases of superposition: from partial to full representation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36455 - Where Root Cause Analysis Fails: A Retrieval-Reranking Decomposition Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36686 - Representable but Unlearned: Encoding Rank and the Interaction-Prediction Floor Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36208 - HeurEvo: Agentic Evolution of Hybrid Solver-Augmented Heuristics for Time-Critical Mathematical Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36303 - When Is Coarse Supervision Worth It? Cost-Aware Learning under Unknown Aggregation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36704 - Safe-by-Design Learning via Energy-based Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36942 - MONOVAB : An Annotated Corpus for Bangla Multi-label Emotion Detection Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2309.15670 - TopoEmbedX: A General Framework for Representation Learning on Topological Domains Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37884 - On State Reduction in Linear Attention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.04852 - RESCUE: Repairing Language Model Errors to Sparse Circuits via Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36813 - Learning Nonlinear Finite Element Solution Operators using Multilayer Perceptrons and Energy Minimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2412.04596 - Fractional State Space Transition for Long Sequence Modeling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36314 - Fair Policy Optimization in Major-Minor Weakly Coupled Markov Decision Processes Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36174 - Quantum Geometry of Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2507.21135 - From Static Policies to Adaptive Priors in Offline Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35880 - Inducing Process Supervision from Outcome-Only Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36641 - Multi-task learning for the automatic grading of enlarged perivascular space burden using MRI Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37387 - Training LLMs to Verbalize Evaluation Awareness Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36316 - Selective Elicitation as a Commercial Influence Channel: A Reproducible Synthetic Shopping-Agent Stress Test Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36614 - Can Representation Learning Decouple from Loss Minimization? Polar Updates Have an Answer Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36240 - PreviewDiff: Multimodal Critic-Guided Search over Diffusion Latents Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36199 - Draft in Parallel, Condition Through Depth: Adjacent Causal Injection for Speculative Decoding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36173 - Early Learning Shapes Later Directions Of Representation Change In Continual Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36081 - Scene-Consistent Illumination Transfer for Inserted Advertising Graphics Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37951 - Introducing the CZAR Loss: A Tailored Objective Function for Financial Log-Return Predictions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36061 - The Camera Inside the Editor: Reading the Implicit Camera of Image Editors with Painted Calibration Patterns Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37732 - Principled Thoughts for Latent Recursive LLM Systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36159 - cktFormer: Transformer-Based Approach for Automated Analog Circuit Design Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36752 - Nonpreemptive Scheduling While Learning Context-Dependent Service Rates Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37660 - Scaling Video Generation for Reasoning: At What Cost? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36599 - GLASS: Global Latent Aggregation with Slot-based Set Decoding for Scalable All-Atom Crystal Generation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37158 - Feedback-Calibrated Protein Optimization with Batch-Aligned Tail Arbitration Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37808 - Transversal Pooling Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36237 - DimGrow: Memory-Efficient Field-level Embedding Dimension Search Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2505.12683 - GARDiff: Graph-Aligned Residual Diffusion for Probabilistic Multivariate Time-Series Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37694 - Probabilistic Symbolic-Distillation Model of Droplet Collision for Spray Simulation at High Ambient Pressures Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37202 - Learned Compression of SAR Phase-History Data: A Rate-Honest Feasibility Study on GOTCHA Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35848 - Evaluating Bounded Autonomy in Regulated Agentic AI: A Diagnostic Harness with Constitutional Rewards, Escalation Labels, and Runtime Governance Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37501 - Adapting Context Compression for Long-Horizon Agents with Counterfactual Continuations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36526 - Are In-Context Images Worth 10 Dimensions? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37659 - PowerZooJax: A JAX-based Power System Benchmark for Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36052 - Learned Queries and Keys Are All You Need: Replacing the Value Projection with Structured Transforms Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36698 - Asymptotic Universal Alignment: A New Alignment Framework via Test-Time Scaling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.08777 - RolloutFaith: Auditing Persistent Internal Interventions in Visual World Model Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36843 - CPATTA: Conformal Supervision Allocation For Active Test-Time Adaptation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2509.25692 - Channel-Dependent State Space Model for Multivariate Time Series Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36453 - PoE-Fuse: Precision-Weighted Expert Fusion for Bi-Temporal Change Understanding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37485 - Quantum Computing for Network Security Classification: Near-Term Classification and Long-Term Memory Efficiency Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36479 - The Signed Geometry of One-Shot Recourse: On-Path Validity and the Signed-Curvature Criterion Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36252 - AdaKerNet: Neural Kernel Decoding for Task-Adaptive Prediction with Multimodal Large Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36368 - Longer Records, Broader Invariance: The Hidden Scaling Problem in Longitudinal Contrastive Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36409 - The Vote Hides the Failure: Aggregation Choice and Noise Robustness in Heart Murmur Detection Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37161 - Safer Content or Firmer Refusals? A Hybrid Perturbation Defense for Alignment under Harmful Fine-tuning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36862 - Explainability from Training with Applications to TCR-Epitope Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36354 - Serverless gossip training of LSTM failure detectors: A matched-protocol comparison with federated, local and centralized learning on NASA C-MAPSS Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35792 - Replication Failure and Trivial Baselines in Road-Level Crash Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35917 - Reference-Guided Machine Unlearning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.11210 - The Universal Classifier for Graph Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36302 - DisCoMBO: Steering Expert-in-the-Loop Black Box Optimization via Distributional Conformance Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36472 - KernelOnet: An Interpretable Neural Operator Based on Kernel Functions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35938 - Procedural Core: A Compact Recurrent Initialization for Vision Transformers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37631 - Self-Supervised Representation Learning as Mutual Information Maximization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.01345 - ImbalancE: Inference-Time Latent Search Against Degree Imbalance in Link Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36996 - Post-Anomaly Detection Inference for Deep SVDD Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37935 - Sharp Convergence and Sampling Trade-offs for Riemannian Diffusion under Nonnegative Ricci Curvature Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36568 - Adapting Linear-Time Architectures for Tabular In-Context Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36337 - It's All Training: A Fully Synthetic Single-Stage Recipe for LLMs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37891 - SemPSG: A Semantic Channel-Aware Foundation Model for Polysomnography Analysis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36619 - Understanding Private Evolution as Learning-Augmented Clustering Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36678 - Intrinsic Associative Memory on Riemannian Manifolds: Curvature, Capacity, and Emergent Modes Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35948 - WUSH-KV: KV Cache Quantization with Data-Adaptive Transforms Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38121 - ReLOBGen: Replayable Limit Order Book Message Generation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35867 - LLMs Learn to Evade Latent Monitors from Prior Feedback Alone Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36490 - CADENCE: A Confidence-Adaptive Dual-Expert Network for Fast and Accurate Time Series Classification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35929 - Unlocking the Critic: Reward-Free Policy Optimization for LLM Post-Training Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37119 - Where the Model Changes Its Mind: Hindsight-Divergence Localization for Efficient Reinforcement Learning with Verifiable Rewards Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36864 - Predictive Geometry of Hidden Trajectories in Transformers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37717 - Guard Models Are Overconfident Where Base Models Are Uncertain Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36477 - SEED: Self-Speculative Decoding via Implicit Encoder-Decoder Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36590 - Equally Good, Yet Different: Benchmarking Rashomon sets in AutoML packages Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36970 - Parallel Tempering for Diffusion-Based Combinatorial Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37323 - Lucid Dreaming for World Models: Learning to Doubt Imagination and Decide by Trust Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37156 - Beyond Keywords: Leveraging Generative LLMs and Label Aggregation to Classify Economic Policy Uncertainty in News Articles Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35856 - Simultaneous Neural Optimal Transport Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37424 - Learning Task Mixtures from Task Affinities: A Probabilistic Graphical Model for Supervised Fine-Tuning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2507.12612 - Cheap and Powerful Tests for Supervised Subspaces: Per-Component Inference for PLS Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36307 - Representational and Functional Robustness to Electrode Montages in EEG Foundation Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36288 - DSWM: Decomposed Spatio-Temporal World Model for Demand-Driven UAV Base Station Repositioning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36845 - Koa-action: Fast and Consistent Structured Decision Making with Generative LLMs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36115 - Data-Free Pruning of Self-Attention Layers in LLMs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.20636 - Dynamic Regret via Discounted-to-Dynamic Reduction with Applications to Curved Losses and Adam Optimizer Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.08372 - Learning from the Gap Between Pass@K and Pass@1 Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35793 - More Programs or More Rolls? Separating Coverage from Specialization in LLM Harnesses Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35873 - Context Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37725 - Interacting particle guidance for sampling reward-tilted generative priors Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37227 - From Seeds to Semantics: Measuring Semantic Accessibility in Deterministic Diffusion Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.06155 - Variational Mixtures and Multi-Marginal Flow Matching: Advancing Statistical Inference with Biological Applications Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36911 - A Sharp Transition in Data Reconstruction under Differential Privacy Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37344 - A Model-Agnostic Physics-Guided Adapter for Few-Shot Transfer of Coastal Flood Prediction Models to Unseen Regions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37565 - TabKD: Tabular Knowledge Distillation through Interaction Diversity of Learned Feature Bins Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.15481 - Efficiently Approximating Attention Is Hard Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37261 - The Default Trap: Rethinking Plan Evaluation in Tool-Using LLM Agents Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36829 - One-Step Next-Latent Prediction Is Not a World Model Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36227 - Engineering Efficient Self-Play Chess: Search, Replay, and Throughput Under Limited Compute Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37447 - Width Expansion as a Method for Class Incremental Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37702 - HyDI: A hybrid Deep Learning-Inductive Logic Programming ensemble for multi-label classification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37740 - Scale Sensitivity in Low-Bit Post-Training Quantization: Curvature of the Quantization Error Landscape Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37416 - Learning Macroscopic Dynamics without Reconstructing Microscopic States Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37392 - CipherGenome: Homomorphic Inference for Genomic Mixture-of-Experts Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35883 - From Checkpoint Variation to Selection Gains in Supervised Fine-Tuning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36569 - Optimal Quantum-Classical Separations for Exact Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38073 - Fooling Algorithms in Non-Stationary Bandits using Belief Inertia Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.05620 - Privy to the Foil: Recasting Value Estimation with a Self-Privileged Critic for RLVR Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37825 - Making Cross-Continental Federated Learning Repeatable with FLIP: a Multi-Application Study Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36001 - Delta-Matching: Closing the Final Gap of Native 8-bit Training for LLMs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37852 - Efficient and Scalable Physics-Guided Fully Convolutional Spatiotemporal Learning for 3D Microstructure Evolution Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36504 - Bregman Consensus Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35930 - Receptive-field-constrained stimulus optimization for human early and intermediate visual cortex Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36391 - Scaling Full Conformal Image Classifiers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37298 - RACE: Relation-Level Counterfactual Explanations for Heterogeneous Graph Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37650 - ReLMem: Learning Recurrent Memory for Longitudinal EHR Modeling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37587 - Stochastic Heavy Ball with Polyak Step Size and Armijo Line Search: A General Convergence Analysis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36668 - Looped Actor: Depth-Recurrent Reasoning Models for Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37432 - Hyperspherical Semantic Trajectory Analysis: Mapping Technological Diffusion across Academic Preprints, Patent Signals, and Compute Scaling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35845 - doPlan: A Variable-Horizon Dataset for Multi-Stage Language-Conditioned Planning in Autonomous Driving Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38028 - Wasserstein Causal Forests for Distribution-Valued Outcomes Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.35898 - The Domain Is a Residue: Adapting Self-Supervised Features, Not Generators Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37330 - Flattening the Connectome Spectrum: A Spectral Filter for FC Induces a Pretraining Target for fMRI Encoders Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37642 - Cross-attention encoding models reveal dynamic spatiotemporal routing across human higher visual cortex Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36366 - Strict-Saddle Landscapes and Multi-Rank Geometry in Low-Tubal-Rank Tensor Sensing Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37865 - EvoMO-SR: Multiobjective LLM-based Evolution of Symbolic Expressions with substructure guidance Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36187 - AutoMark: Enabling Autoresearch to Discover Better LLM Watermarks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37310 - Correct, Don't Delete: Mitigating Emergent Misalignment with Corrective Supervision Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37624 - BASE: Batch-Aware Selection of Experts Using Predicted Removal Error for Efficient MoE Decoding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.36222 - Almost Human, Except When It Matters: VoxParity and the Decisions a Voice Should Change Source: arxiv-cs-lg Topic: AI Research (+AI, JavaScript, React) URL: https://arxiv.org/abs/2609.35922 - SelfSearch: Reward-Free Search for Self-Improving Agents Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.37968 - Do Self-Evolving Skills Generalize to Held-Out Tasks? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39148 - Persona and Persuasive Framing in AI Voice Agents: A $2\times2$ Field Experiment with Children Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38782 - Social Choice Foundations for Simulation-Augmented Generation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38287 - EviRover: Reinforcing Agentic Perception Beyond a Glance Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.40230 - EvoSteer: Online Self-Evolving Graph Orchestration via Reference-Anchored Credit Assignment Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38661 - Disentangling Self-Distillation: Measuring and Modeling Acquisition and Retention Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39494 - EDGC: Entropy-driven Dynamic Gradient Compression for Efficient LLM Training Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.10333 - Travel Time Prediction in Supply Chain Management Using Machine Learning Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38190 - WinoTS: Wavelet-based Self-Distillation for Time Series Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39337 - Trust Is Not a Score: Runtime Assurance Contracts for High-Risk AI Agents Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39717 - Aligned Data Can Induce Misalignment via Context Confusion Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38379 - Can Computation from Earlier Problems Help LLMs Solve New Ones? Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.39394 - Sparse-WAM: Accelerating World Action Models via Action-Guided Sparse Imagination Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38984 - Group Preference Collapse in Personalized Multimodal Large Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.22603 - Signal-Routed Temperature Scaling: Low-Capacity Risk-Conditioned Calibration for Small Validation Budgets Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.38936