AI News

How Humans Learn Unfamiliar Worlds: A Think-Aloud Dataset on ARC-AGI-3 (arxiv.org)

Deep Learning-Based Prediction of HER2 Status from Whole-Brain MRI in Breast Cancer Brain Metastases (arxiv.org)

Direction-Aware Token Credit Assignment with Reward-Model Gradients (arxiv.org)

On the Identifiability of Hybrid Deep Generative Models: Meta-Learning as a Solution (arxiv.org)

DT-JEPA: Learning to Predict Future Representations for Decision-Making (arxiv.org)

WL-SeqVAE: Sequence-Based Graph Generation Using Variational Autoencoders and Weisfeiler-Leman Colors (arxiv.org)

CaliMem: Search-Efficient Data Agent via Calibrated Memory (arxiv.org)

What Effort Level Should I Choose? On the Economics of Reasoning Effort in LLM Agents (arxiv.org)

From Outranking Flow Scores to Linear Programs: Group Decision Making in PROMETHEE II (arxiv.org)

CTXAgent - Self-Controlled Context Management of LLM Agents (arxiv.org)

Matched Capabilities, Different Behaviors: A Controlled Study of MCP and CLI Interfaces for Language Agents (arxiv.org)

The Standard Deviation in VICReg Buys a Safety Margin (arxiv.org)

Earth embeddings are decomposable records of the land surface (arxiv.org)

Interpreting Supervised Pairwise Graph Representations: Local Evidence, Confounding, and Provable Post-hoc Correction (arxiv.org)

Trajectory Matching: Discrete Flow Matching with Structured Actions (arxiv.org)

From Parameter Rebasin to Functional Transfer Across Heterogeneous Transformers (arxiv.org)

MultiEvoSkill: Structure-Aware Multi-Agent Skill Evolution for Generalizable Skill Artifacts (arxiv.org)

Variational Cognitive Memory Abstraction for Long-Horizon Vision-Language-Action Models (arxiv.org)

MatOS: An Agent Operating System for Materials Research (arxiv.org)

R2A-Sales: A Policy-Atom-Matched Benchmark for Rule-to-Action Gaps in Goal-Directed Sales Agents (arxiv.org)

Task Success Is Not Enough: Side-effect-Aware Evaluation of Tool-Using Language Model Agents (arxiv.org)

Shared Is Not Enough: Predictor Width and Factor Readability Under Masked Latent Prediction (arxiv.org)

Principled Nonlinear Concept Erasure and Steering (arxiv.org)

Amortized Thompson Sampling and Efficient Information-Directed Sampling for Simulator-Based Design (arxiv.org)

BAARC: Uncertainty-Driven Adaptive Partial Spectral Measurement to Full Reconstruction for Material Classification (arxiv.org)

DRIVE: Fine-tuning via Data Contribution- and Diversity-aware Weighting with Prior Regularization (arxiv.org)

Can LLMs Trust Appropriately? Evaluating Calibrated Reliance on Imperfect Partners (arxiv.org)

From Document Layout to Causal Topology: An End-to-End Architecture for Temporal Causal Reasoning (arxiv.org)

Fed-Universe: A Semantic-Geometric-Topological-Human (S-G-T-H) Stack for Negotiated Alignment in Federated Systems (arxiv.org)

WebViSE: Scaling Visual Design Intelligence from the Open Web for LLM Website Generation (arxiv.org)

OS-Wizard: Scaling Computer-Using Agents to Specialized Software via Knowledge-Guided Exploration (arxiv.org)

Accordion: Elastic-Depth Inference via Fixed-Prefix Layer Looping (arxiv.org)

A curated multi-endpoint drug-protein bioactivity dataset with construct-aware protein targets (arxiv.org)

Online Continual Learning in Edge-Cloud Applications (arxiv.org)

Sharpness-aware Orthogonalized Momentum Algorithms (arxiv.org)

DualGSL: Prior-Guided Graph Structure Learning with Node-Adaptive Prediction Fusion (arxiv.org)

ReGE: Stronger Agents through Experience Synthesis (arxiv.org)

Numina Fuse: Interactive Autoformalization in Lean 4 (arxiv.org)

Build, Don't Play: Self-Improving LLM Research Agents Develop a Hearthstone Battlegrounds AI That Rivals Top Players (arxiv.org)

EnvLunch: The Environment as a Free Lunch for Learning to Interact (arxiv.org)

Detection and Validation of Hidden Subgroup Disparities in Clinical AI (arxiv.org)

CA-HNM: Constraint-Aware Hard Negative Mining for Dense Retrieval with Structured Domain Models (arxiv.org)

AutoPRL: Automated Path Representation Learning via a Learnable View Generator (arxiv.org)

From Single Spins to Ensembles: A Calibrated Symbolic Regression Pipeline for NV Spectroscopy (arxiv.org)

Belief-Conditioned Multi-Agent Actor Critic for Decentralized LLM Collaboration (arxiv.org)

AskGLP1: A Knowledge Graph LLM Agent for Exploring GLP-1 Experience from Online Patient Narratives (arxiv.org)

Sparse Recovery Bird's-Eye View Model Soup for Robust Camera-Based 3D Object Detection (arxiv.org)

SVD-Attention: Efficiently Analyzing Attention Blocks in Transformers (arxiv.org)

The Target Agent Is Its Own Best Critic but a Poor Proposer: Self-Improved Red Teaming (arxiv.org)

Sustainability of TSFM Deployment and Pretraining: From Carbon Footprint Analysis to Mitigation via Pruning (arxiv.org)