AI News

APEx: Distillation of Agent Procedural Experience for Adaptive Deep Research Question Answering (arxiv.org)

LLM-as-a-Judge Is Not an Oracle: Why Self-Improving Agents Need Deterministic Guardrails (arxiv.org)

Task-Level Natural Language Priors as Learning Signals for Low-Resource LLM Training (arxiv.org)

Propose to Learn, Learn to Propose: Evaluability-Aware Assistance under Bounded Rationality (arxiv.org)

PGPO: Potential-Guided Policy Optimization for Multi-Turn Agentic Tasks (arxiv.org)

PhoenixNest-Video: Evidence-Grounded Multimodal Agent Framework for Automated Video Interview Assessment (arxiv.org)

SkillGLoW: Procedural-Family Skill Consolidation for Self-Improving Agents on Long-Horizon Task Streams (arxiv.org)

PEARL: Path-Entity Aligned Relational Learning with Contextual Subgraphs for Inductive Knowledge Graph Completion (arxiv.org)

ASCII Attack: Recontextualising Harmful Requests as Artistic Critique in Large Language Models (arxiv.org)

Quantum MeanFlow: single-shot generative sampling on NISQ hardware (arxiv.org)

Breadth Beats Depth: Improving GCG-Based Jailbreak Optimization with Breadth-Oriented Suffix Search (arxiv.org)

GenCAR: Generative Counterfactual Alignment with Risk-Controlled Selection for Out-of-Distribution Recommendation (arxiv.org)

A Power Law in Logarithm's Clothing: On the Scalability of Graph-Based Vector Search (arxiv.org)

SoK: Where Do Flow Labels Come From? Auditing Label Provenance in Encrypted Traffic Benchmarks (arxiv.org)

Disease Burden over Skin Tone: Decomposing the Dermatology-AI Generalization Gap (arxiv.org)

WeaveMark: Robust and Scalable Multi-bit LLM Watermarking via Coded Payload Spreading (arxiv.org)

IDEEA: training-free Input-Dependent stEEring via Activation cluster matching (arxiv.org)

DocHop: Benchmarking Out-of-domain Multi-hop Reasoning in Information-Dense Documents (arxiv.org)

Perceptually Regularized Diffusion Model for Image Super-Resolution (arxiv.org)

Morphology signal in whole slide image foundation models can automatically triage slides (arxiv.org)

Post-Training Ternarization of Qwen3-4B Capability, Effective Bit Budget, Storage Compression, and Deployment (arxiv.org)

Network-Aware Forecasting on Wireless Access Points (arxiv.org)

Pushing Forward Multi-Secret-Key Homomorphic Encryption for Private Average Aggregation (arxiv.org)

Sparse Readout Prism: Explaining Logit-Lens Scores in Features Instead of Tokens (arxiv.org)

Basin Geometry and Reliable Recall of Dynamical Memories in Reservoir Computing (arxiv.org)

Latent unified smooth Hamiltonians for excited state chemistry (arxiv.org)

Interpretable Symptom Vectors for Depression in a Large Language Model (arxiv.org)

Reinforcement learning to choose optimizers (arxiv.org)

Ten Architectures, One Error: Shared Failure Modes in Hyperspectral Classification under Spatially Disjoint Evaluation (arxiv.org)

Harness Engineering in LLM Tool Use via Agent-Native Reusable Tool Primitives (arxiv.org)

Hearing the Whispers: Black-Box Membership Inference Attacks on Finetuned TTS Models (arxiv.org)

FairLens: Benchmarking Fairness in Vision-Language Models for High-Stakes Decision-Making (arxiv.org)

FORGE: Forward-Only Test-Time Adaptation for Integer-Only Vision Models on Microcontrollers (arxiv.org)

Private Computation Space: Experience with Trusted Multi-Cluster Federated Learning for Agriculture (arxiv.org)

Context Inference Attacks Without Jailbreaks (arxiv.org)

SocialBuddy: Tailoring Search Agent for Social Scenarios (arxiv.org)

Omega-N: Interpretable Structural Node Descriptors and Their Applicability Domain (arxiv.org)

Marginal Expected Revenue for Jointly Ranking Auction and Fixed-Price Listings in E-Commerce Sponsored Search (arxiv.org)

Random Forest-Informed Cellular Automaton for Large-Scale Wildfire Spread Modelling (arxiv.org)

PRISM: An Agentic Multi-Model Architecture for Proactive Safety in Autonomous Transportation Systems (arxiv.org)

RecEvolve: A Knowledge-Driven Autonomous Agent System for Recommender Systems (arxiv.org)

When Literature Data Mislead Artificial Intelligence in Materials Discovery (arxiv.org)

Graph Machine: Towards Better Pretraining via Edges (arxiv.org)

The Implications of Linguistic Illegibility for LLM Security (arxiv.org)

Post-Training Language Models for Gold-Medal Performance in Coding Competitions (arxiv.org)

UE5M3 FP4 Block Scaling for Stable Language Model Pretraining (arxiv.org)

Cliff: Learning Process Rewards from the First Mistake (arxiv.org)

A Study of Conditional Diffusion Models for Open-Loop Control under Dry Friction and Stiction (arxiv.org)

Shiva-DiT: Residual-Based Differentiable Top-$k$ Selection for Efficient Diffusion Transformers (arxiv.org)

The Ensemble Kalman Inversion Race (arxiv.org)