AI News

RASER: Recoverability-Aware Selective Escalation Router for Multi-Hop Question Answering (arxiv.org)

Iteris: Agentic Research Loops for Computational Mathematics (arxiv.org)

Temporal Motif Signatures for Temporal Graph Neural Networks (arxiv.org)

Fairness in two-player zero-sum games with bandit feedback (arxiv.org)

When Data Is Scarce: Scaling Sparse Language Models with Repeated Training (arxiv.org)

Lagrangian Perturbation Diffusion Steering: Latent Reinforcement Learning for Generative Policies (arxiv.org)

STARFISH: faST Accuracy Recovery in pruned networks From Internal State Healing (arxiv.org)

MCP-Persona: Benchmarking LLM Agents on Real-World Personal Applications via Environment Simulation (arxiv.org)

AGENTCL: Toward Rigorous Evaluation of Continual Learning in Language Agents (arxiv.org)

Beyond One-shot: AI Agents for Learning in Field Experiments (arxiv.org)

HLL: Can Agents Cross Humanity's Last Line of Verification? (arxiv.org)

Bridging the Sim-to-Real Gap in Semiconductor Visual Program Synthesis via Input Binarization (arxiv.org)

From Reward-Free Representations to Preferences: Rethinking Offline Preference-Based Reinforcement Learning (arxiv.org)

A Per-Component Diagnostic Protocol for Neural HJB-PIDE Solvers under Control-Dependent L\'evy Jumps (arxiv.org)

HASTE: Hardware-Aware Dynamic Sparse Training for Large Output Spaces (arxiv.org)

LeAP: Learnable Adaptive Permutation for Feature Selection in Heterogeneous and Sparse Recommender Systems (arxiv.org)

Soft-NBCE: Entropy-Weighted Chunk Fusion for Long-Context (arxiv.org)

A Fiber Criterion for Representation Identifiability in Supervised Learning (arxiv.org)

AgentPLM: Agentic Protein Language Models with Reasoning-Augmented Decoding for Protein Sequence Design (arxiv.org)

A Mathematical Conflict Framework for Contextual Data Modulation (arxiv.org)

Spatial Representation Learning Beyond Pixels: Unifying Raster Data and Vector Semantics for Human-Centric Geospatial Foundation Models (arxiv.org)

Harness-1: Reinforcement Learning for Search Agents with State-Externalizing Harnesses (arxiv.org)

COMAP: Co-Evolving World Models and Agent Policies for LLM Agents (arxiv.org)

PaCX-MAE: Physiology-Augmented Chest X-Ray Masked Autoencoder (arxiv.org)

MViewRouter: Internalizing Geometric Equivariance via Multi-view Alternating Attention for Combinatorial Routing (arxiv.org)

Decision-Focused On-Policy Learning for Contextual Linear Optimization with Partial Feedback (arxiv.org)

ThinkSwitch: Context Distillation with LoRA and Weight Interpolation for Specific-Purpose Reasoning Tasks (arxiv.org)

Non-Vacuous Certification of Transport MCMC via Oscillation-Controlled Normalizing Flows (arxiv.org)

Improving Hospital Process Management through Process Mining: A Case Study on COVID-19 Clinical Pathways (arxiv.org)

MOC: Multi-Order Communication in LLM-based Multi-Agent Systems (arxiv.org)

SIRI: Self-Internalizing Reinforcement Learning with Intrinsic Skills for LLM Agent Training (arxiv.org)

Coordination Graphs for Constrained Multi-Agent Reinforcement Learning (arxiv.org)

Repair Before Veto: Repair-Augmented Constraint Learning for Contextual Decisions (arxiv.org)

Interaction-Limited Safe Continuous-Time RL for Dynamical Medical Treatment (arxiv.org)

Plausibility Is Not Prediction: Contrastive Evidence for LLM-Based Cellular Perturbation Reasoning (arxiv.org)

OPD+: Rethinking the Advantage Design for On-Policy Distillation (arxiv.org)

MedGym:A Unified Continuous-Time Benchmark for Dynamic Medical Treatment Reinforcement Learning (arxiv.org)

Strong Stochastic Flow Maps (arxiv.org)

POIROT: Interrogating Agents for Failure Detection in Multi-Agent Systems (arxiv.org)

CEON: Circular Economy Ontology Network (arxiv.org)

From Capability Models to Automated Planning: An AAS-Native Approach for Automatic PDDL Generation (arxiv.org)

LLM-Evolved Pattern Generators for Optimal Classical Planning (arxiv.org)

S3TS: Stochastic Scenario-Structured Tree Search for Advanced Planning Under Uncertainty (arxiv.org)

Beyond Task-Agnostic: Task-Aware Grouping for Communication-Efficient Multi-Task MoE Inference (arxiv.org)

Trust Functions: Near-Lossless Weak-to-Strong Generalization by Learning When to Trust the Weak Teacher (arxiv.org)

Data Enrichment for Symbolic Regression Using Diffusion Models (arxiv.org)

Profiling Privacy Preservation Against Gradient Inversion Attacks in Tabular Federated Learning (arxiv.org)

CryoProt: A Protein Pretraining Framework with Cross-Box Interactions on Cryo-EM Density Maps (arxiv.org)

BADGER: Bridging Agentic and Deterministic Evaluation for Generative Enterprise Reasoning (arxiv.org)

eMoT: evolving Memory-of-Thought via Symbolic Anchoring and Memory Corrosion (arxiv.org)