AI News

POSEIDON: Scaling Training Environments for Scientific Agents from Code Repositories (arxiv.org)

Tradeoff-Aware Territory Design: A Multi-Objective Decision System for Enterprise Customer-to-Seller Allocation (arxiv.org)

AgentTrace: Execution-Level Fingerprinting of Black-Box LLM Agents (arxiv.org)

ALLMTS: Automated LLM-based Medical Triage System (arxiv.org)

Adaptive Control of Scientific Instruments with Nonstationary Responses through Multimodal Agents (arxiv.org)

MBABench-2.0: Evaluating AI Agents on Constructing Financial Spreadsheets Beyond Simple Verifiability (arxiv.org)

MusicTwins: Training Audio-Language Models to Listen like Their Persona (arxiv.org)

REACT: Representational Early Analysis and Control of Chain-of-Thought Trajectories (arxiv.org)

TTT in VLA Models: Exploring Training with Human Demonstrations (arxiv.org)

MG-REPA: Multi-Granularity Representation Alignment for Generative Target Speaker Extraction (arxiv.org)

HyScale: Scaling Long-Context Training for Learned Sparse Attention (arxiv.org)

Do AI-for-Science Agents Preserve Model Safety? (arxiv.org)

OpenHome: From Perception to Insight in Smart-Home Camera Videos (arxiv.org)

Knowing When to Stop: The Evidence-Action Gap in Coding Agents (arxiv.org)

Controlled Evaluation of EEG Foundation Models: Random Initialisation, Band-Power and Validation-Selected Probes (arxiv.org)

Learning Structured Sparsity for Large Models: Bridging Model Compressibility and Efficient Inference (arxiv.org)

Agentic AI Defense: Decode-Time Provenance Binding for MCP (arxiv.org)

Spectra: Multi-Scale Implicit Decomposition for Deepfake Detection (arxiv.org)

Tampered ML: Evaluating LLM Judgment of Machine Learning Research (arxiv.org)

PAIR-SAFE: A Paired-Agent Approach for Runtime Auditing and Refining AI-Mediated Mental Health Support (arxiv.org)

PrefO2O: Zero-Shot Preference-Based Reinforcement Learning via Offline-to-Online Latent Preference Transfer (arxiv.org)

Minimum Description Length Unsupervisedly Separates Law-Grade from Statistical-Grade Knowledge in a Self-Improving Science Agent - and Reveals 86% of Its Own Discoveries Are Statistical (arxiv.org)

Rethinking Multi-View Fusion for Urban Representation Learning (arxiv.org)

What the Query Remembers: Query Abstraction Governs Lexical and Dense Retrieval of Classical Arabic Poetry (arxiv.org)

Benchmarks Are Surprisingly Compaction-Friendly (arxiv.org)

Searching with Sparse Histories: Cross-Task Architecture Search from Incomplete Search Logs and Zero-Cost Proxies (arxiv.org)

Beyond Pattern Matching in 3D MLLMs: Part-Level Functional Reasoning for Generalizable Recognition (arxiv.org)

A Mathematical Theory of the Jacobian Lens (arxiv.org)

RSI-Opt: Recursive Self-Improvement Optimizer Agents For Model Training (arxiv.org)

Anti-coherence Regularized Absolute Scaling for Out-of-Distribution Detection (arxiv.org)

When Agents Fail: Model or Harness--Who's to Blame? (arxiv.org)

RobustNCL: Robustifying Neuromorphic Continual Learning under Transient Faults with Lightweight Detect-and-Cure Strategy (arxiv.org)

Heterogeneous Space Data Centers: A Key Paradigm for Future Non-Terrestrial and In-Orbit Services (arxiv.org)

Representation Learning via Composite Subspace Flow Matching (arxiv.org)

Dual-Adaptive Z-Score PPO: Dynamic Clipping Using Action Confidence and Normalized Advantage Magnitude (arxiv.org)

Tokamak-SI: Interpreting Tokamak Discharges with Labeled Catalogs (arxiv.org)

"Not" is Not Lost in Translation: A Causal Analysis of Negation Representations in Multilingual LLMs (arxiv.org)

Concept Acupuncturist: An Efficient and Precise Concept Erasure Method for Diffusion Models via Sparse Autoencoders (arxiv.org)

Complementing DINO Features with Image Structure for Part Discovery (arxiv.org)

Uni-Stream: A Unified Streaming Speech Language Model for Speech Synthesis and Recognition (arxiv.org)

Do VLAs Know They Will Fail? Interpretable Success and Failure Patterns in Physical AI Systems (arxiv.org)

End-to-End Differentiable Diffusion Conditioning for Physics-Informed Optimization (arxiv.org)

Generalized green and red concepts for artificial intelligence, big data, and information and communication technologies: Multidimensional framework and countermeasures (arxiv.org)

Reuse, Rewrite, or Reject: Evidence-Constrained Trace Reuse for Repeated Mobile GUI Control (arxiv.org)

FSVL: Taxonomic Fast-Slow Vision-to-Language Collaborative Framework for Interpretable Confusable Fish Species Identification and Knowledge Generation in the Middle Yangtze River (arxiv.org)

Machine-Readable Operator Manuals over the ISO 11783 Network for Offline, Multilingual LLM Assistance (arxiv.org)

Autoregressive Multimodal Anomaly Detection with Large Language Models for Microservice Systems (arxiv.org)

Beyond Data Schedule: Fast LLM Pretraining on Data Mixtures by Loss-Reweighting (arxiv.org)

ChoreoStudio: Planning What, When, and Who in Music-Driven Group Choreography (arxiv.org)

CASE: Confusion-Aware Skill Evolution for Frozen Multimodal Emotion Models (arxiv.org)