AI News

EvoSim: Learning to Model, Modeling to Learn (arxiv.org)

ReCast: Attribution-Oriented Step Representation Learning for LLM-Based Agent Systems (arxiv.org)

TokenBank: Financial Infrastructure for AI Services (arxiv.org)

Mine Odyssey: Benchmarking Spatial Agentic Intelligence in the Wild (arxiv.org)

F$^3$NO: Frequency-Decomposed Finite-Time Flow-map Neural Operators with Cross-Scale Conditioning (arxiv.org)

HAND: A Biologically-Inspired Activation Function that Improves Generalisation and Sample Efficiency in Image Classification (arxiv.org)

Open-Vocabulary Audio-Visual Event Localization via Complex-Valued Fusion (arxiv.org)

Learning qBIC Resonances across Metasurface Families in Dielectric Fourier Space (arxiv.org)

Evaluating Local Language Model Agents for Reproducible Data Engineering: An Empirical Software Engineering Study of Mobility Workflows (arxiv.org)

Finsler Flow Matching: Dynamics-Aware Geodesic Interpolation for Single-Snapshot Trajectory Inference (arxiv.org)

DivMoE: Fine-Grained MoE Upcycling via Cross-Domain Expert Composition (arxiv.org)

Conditional Residual Prediction: Improving Autoregressive Video Diffusion without a Bidirectional Teacher (arxiv.org)

MedBenchAgent: Towards Systematic Automation of Medical VLM Benchmark Construction (arxiv.org)

Who Verifies the Verifier? Co-Evolving Inspectable Graders with Self-Improving Agents (arxiv.org)

Closed-loop evaluation of LLM agents for embedded software development (arxiv.org)

DuplexAgent-RSI: Recursive Harness Improvement for Full-Duplex Voice Agent Collaboration (arxiv.org)

ORDO: Operation-level Round-aware Dynamic Ordering for MIP Presolve (arxiv.org)

Open-ended Scientific Discovery with Possibilistic Reasoning (arxiv.org)

TaReD: Tool-Aware Recursive Decomposition for Long-Horizon Tasks (arxiv.org)

LLM-IDEA: Identifiability-Driven Experimental Agent for Autonomous Discovery of Mechanistic World Models (arxiv.org)

BioBigBird: A Sparse Attention Model for Long-Range Dependency Processing in Biomedical Text (arxiv.org)

Causal-fate dynamics of unrealized influence (arxiv.org)

Refinement as a Service: Algorithmic Predictor Refinement (arxiv.org)

WAM-Cache: Staleness-Bounded KV Reuse for Efficient World Action Models (arxiv.org)

Optimizing Large Language Models with Chained LMOs (arxiv.org)

Harness Compilation: Which Decisions Should a Small Vision-Language Model Keep? (arxiv.org)

Internalizer: Portable Context-to-Parameter Mapping for Very Large Language Models (arxiv.org)

AliO: Output Alignment Matters in Long-Term Time Series Forecasing (arxiv.org)

What to Admit and How to Present: Governing Persistent Memory in LLM Agents (arxiv.org)

Social Pain Disrupts Emotion-Action Brain-State Dynamics in Adolescents with Non-Suicidal Self-Injury (arxiv.org)

Balancing Reference Guidance and Free Generation in Trajectory Rollouts for Reasoning RL (arxiv.org)

When Interfaces Speak: Data-Aware Generative UI Harness for Active Interaction (arxiv.org)

OpenProblemBench: Benchmarking AI on Open Problems in the Foundational Theoretical Sciences (arxiv.org)

Beyond Imitation: A Framework and Benchmark for LLM-Assisted Peer Review (arxiv.org)

AgentHorizon: Evaluating Agentic Judges for Long-Horizon Computer-Use Tasks (arxiv.org)

Distillation for Incrimination and Distillation for Capabilities (arxiv.org)

Curating Always-Loaded Context for LLM Agents: A Capacitated Assortment Model with Censored Feedback (arxiv.org)

How Narrative Wrapping Affects LLM Refusal: A Cross-Language Benchmark and Defense (arxiv.org)

Teaching PPG How not Who: Fixed-Effects Distillation from ECG (arxiv.org)

Sample-Efficiency of Kolmogorov-Arnold Networks (arxiv.org)

Ranking Prior Alignment for Credit Risk Modeling: When Do External Priors Matter? (arxiv.org)

ActiveMedAgent: Cost-Aware Trajectory Learning for Multimodal Medical Diagnosis (arxiv.org)

PRAXIS: Learning Dynamics of Self-Improving Models with Symbolic Archives (arxiv.org)

Spectral Weight Decay: Inducing Low-Rank Structure in Neural Network Weights (arxiv.org)

Learning to Plan by Looking Back: Hindsight Hierarchies for Training Reasoning Models (arxiv.org)

Learning How to Search for Plans with Exponentially Less Space (arxiv.org)

BEANS-Next and ROOTS: Broadening Audio-Language Capabilities for Bioacoustics (arxiv.org)

StoreBench: A Live-Commerce Environment for Evaluating and Training Autonomous Operator Agents (arxiv.org)

Reading the Room: Foundations, Design, and Challenges of Normative Competence in LLMs (arxiv.org)

Self-Supervised Keyframe Discovery for Horizon-Invariant Behavior Cloning (arxiv.org)