AI News

From Outcomes to Steps: Reliable Success-Aware Critic Learning for Long-Horizon Language Agents (arxiv.org)

Learning Inter-Graph Interactions Between Heterogeneous Substructures of Chemical Systems (arxiv.org)

Auditing the Product, Not the Recursion: The Evaluation Evidence Behind Recursive Self-Improvement Claims (arxiv.org)

Bridging Digital Deception and Physical Fidelity: A Spatio-Temporal Attack on Cognitive Radars (arxiv.org)

HypTI: Hyperbolic Geometry-Aware Multimodal Feature Fusion Framework for Dementia Detection Based on Tabular-Image Data (arxiv.org)

From Comparison to Selection: Enhancing Cross-Scale Generalization of Pointer Networks Trained with Reinforcement Learning for Capacitated Drone Delivery Routing (arxiv.org)

Quantum-JEPA: Quantum Born Priors for Latent-Action JEPA World Models (arxiv.org)

ACE-Compression: An Ability-Centric Evaluation of Compressed Large Language Models (arxiv.org)

MMEnterprise-Bench: Evaluating Enterprise Agents over Evolving Multimodal Records (arxiv.org)

Two Is Enough: Binary Classification Priors Suffice for Multi-Class In-Context Tabular Prediction (arxiv.org)

TRCM: Topology-Aware Route-Consistency Modeling for Robust Neural Vehicle Routing (arxiv.org)

AnchorRM: Parallel Intra-Batch State Reconstruction with Anchor-Based Residual Memory for Temporal Graph Neural Networks (arxiv.org)

Can Native Safety Travel? Reusing Safety Pretraining Across Models via On-Policy Distillation (arxiv.org)

Automatic Differentiation with Constraints in Forward and Inverse Problems (arxiv.org)

Physical Operating-Condition-Guided Vision-Language Matching for Cavitation Diagnosis in Industrial Pipeline Systems (arxiv.org)

COMPASS: Routing via Capability-Oriented Model Profiling with Adaptive Search, Selection, and Tool Use (arxiv.org)

Privacy-preserving inference on hosted models: a learned input-embedding transform for frozen Qwen3.8-27B (arxiv.org)

LatBridge: Efficient and Stable Model Replacement in Latent-Communication Multi-Agent Systems (arxiv.org)

Lower Bounds for Smooth Nonconvex Zeroth-Order Optimization: The Price of Determinism under Low Curvature Trace (arxiv.org)

Timing-Aware and Interpretable Self-Attention for Event-Based Wearable Task Transition Detection (arxiv.org)

BEACON: Bridging Multimodal Semantics and Asynchronous Propagation for Cloud System Anomaly Detection (arxiv.org)

Self-Evolving Agents through Privileged Retrospective Distillation (arxiv.org)

Orthogonal Boundary Unlearning for Dual-Use Concepts in Large Language Models (arxiv.org)

When Parameter Gradient Flows Close in Representation Space (arxiv.org)

Training-Free Continual Learning for Multimodal Forecasting via Prompt Evolution (arxiv.org)

Projected Distillation: On Learning to Solve Hard Problems from Close, Correct Targets (arxiv.org)

Toward Agentic AI for the 6G Physical Layer: From Algorithm Discovery to Verified Baseband Silicon (arxiv.org)

Decision Sufficiency of Learned Representation Interfaces for Safe Generative Control (arxiv.org)

Visual Task Modeling and Automated Planning for Humanoid Robots in Education of Individuals with Autism Spectrum Disorder (arxiv.org)

Model Exposure on One Device Endangers Others: Rethinking Edge Neural Network Security across Devices (arxiv.org)

Mind Your Boundaries: Boundary-Aware Model Obfuscation against Cross-node Attack (arxiv.org)

Joint Constraint Aggregation for Qualifier-Sensitive Retrieval-Augmented QA (arxiv.org)

Cross-Lingual Reusability of LLM Pipelines: Financial Extraction, Multilingual Reasoning, and Semantic Screening (arxiv.org)

Robustness of Internal Signals for Hallucination Detection in Vision-Language Models (arxiv.org)

Deficit-Driven Stratified Acquisition for Building STRATA-S12: An Open SAR-Optical Corpus for Self-Supervised Learning (arxiv.org)

Learning Reachability-Based Safety Representations for Real-Time Robot Navigation in Unknown and Dynamic Environments (arxiv.org)

CountVideoBench: A Benchmark to Evaluate Prediction Biases on the Object Counting task by Video-Language Models (arxiv.org)

ABACUS: A Benchmark for Image Editing Models to Assess Object Category and Counting Under Scene Conservation (arxiv.org)

Disclosure Begets Disclosure: A Self-Reinforcing Privacy Loop in LLM Agents with Persistent Memory (arxiv.org)

Adaptive Replay Refinement: Allocating Test-Time Compute Between Revision and Resampling (arxiv.org)

Local Work, Global Progress: A Convergence Theory of AI Agent Collaboration (arxiv.org)

BR-MoE: Budgeted Semantic Release for Privacy-Preserving Traffic Analytics (arxiv.org)

3D-MoE: Complementarity-Guided Hardware-Software Co-Design for MoE Decode on 3D-Stacked NMP (arxiv.org)

GRaSP: Latency-Aware Optimization for Geo-Distributed Retrieval-Augmented Generation (arxiv.org)

How Intervention History Changes AI Agent Sabotage on Subsequent Tasks (arxiv.org)

Convergent Performance, Divergent Mechanisms in LoRA-Adapted Vision Foundation Models for Brain MRI (arxiv.org)

ExeBench-FHD: Addressing Fidelity Issues in Decompiler Evaluation (arxiv.org)

Attractive Faces Skew Text-to-Image Reward Models and Leaderboards (arxiv.org)

Offline Makes Online Better: Calibrated Cache Compression for Video Generation (arxiv.org)

Multilingual Brain Decoding Reveals Shared Cortical Semantic Geometry (arxiv.org)