AI News

Steertuning Policies from a Few Human Videos (arxiv.org)

UniW(A)M: A Unified Mixture-of-Experts Framework for World and World-Action Models (arxiv.org)

GeoReader: Learning Latent Geometry for Spatial Reasoning in Vision-Language Models (arxiv.org)

RadarGSOcc: Doppler-Aware 3D Semantic Occupancy Prediction Using 4D Radar Tensor with Geometric-Semantic Feature Extraction (arxiv.org)

ProAS: A Progress-Aware Advantage Shaping Framework for Multi-Turn Agentic Reinforcement Learning (arxiv.org)

From Relation-Specific Heads to Shared Interventions: Contextual Entrainment in Language Models (arxiv.org)

SMI: Ontology-Grounded Semantic IDs for Hierarchy-Aware Medical Concept Representation (arxiv.org)

Known-Voice Cloning and the Limits of Personality Transfer for Character Design (arxiv.org)

Kwai Summary Attention V2 Technical Report (arxiv.org)

AMemOps: Evaluating Memory Updates in Long Audio Streams (arxiv.org)

Tiny Recursive Message Passing (arxiv.org)

ST-GRU PredRNN-V2: A GRU-Based Reformulation of PredRNN-V2 for Video Prediction (arxiv.org)

Unlocking Token Rewards via Training-Free Reward Attribution (arxiv.org)

FALLNEST: Dual-Domain Dataset Condensation of Imbalanced IMU Data for Wearable Fall Detection (arxiv.org)

LiFUSE: Frequency-Spatial Unified Self-Supervised Pretraining for Liver Disease Analysis on Non-contrast CT (arxiv.org)

Heading Hierarchy Reconstruction for Complex and Long Documents with Level-Wise Multimodal Memory (arxiv.org)

Holding Agents Accountable with Verifiable Agent Insurance (arxiv.org)

A Controlled Evaluation of Curvature Inspired Scalar Feature Scaling for Domain Generalization (arxiv.org)

TextualRL: Textual Reinforcement Learning for Context Management (arxiv.org)

Reset Early, Reset Often, Eliminate Models (arxiv.org)

How Reliable Are Large Language Models for Soil Organic Carbon Assessment? An Australian Benchmark (arxiv.org)

Virtual Head Attention (arxiv.org)

FETALFusion: Resolution-Aware State-Space Modeling for Multi-Domain Fetal Head Segmentation (arxiv.org)

Benchmarking and Training Reward Models for Cross-Source Evidence Grounding in Search Agents (arxiv.org)

Empowering Edge Agents: A Systematic Analysis of Edge-Cloud Collaboration in LLM Agent Execution (arxiv.org)

On Making $\mathrm{SE}(2)$-Invariant Networks Optimal (arxiv.org)

Same Attack, Different Judge: A First Controlled Adversarial Comparison of Jev, LLM Judges, and Purpose-Built Safety Classifiers (arxiv.org)

Learning from fracture: Teaching language models to design through physical experience (arxiv.org)

EmoXpt: Analyzing Emotional Variances in Human Comments and LLM-Generated Responses (arxiv.org)

GIFT-Eval-CTX: A Benchmark For Contextual Forecasting Model Evaluation (arxiv.org)

Posterior-Averaged Bellman Flow Matching for Distributional Critics (arxiv.org)

Efficiently Learning To Reason or Not to Reason: Root-token Policy Optimization for Adaptive Thinking (arxiv.org)

Gaussian Process Bandits for AI System Optimization with Selectively Verified Feedback (arxiv.org)

PaceReasoner: Label-Free Learning to Answer Briefly from Long Reasoning (arxiv.org)

Degrees of Freedom in Multi-Agent Coordination: Between Centralized and Decentralized Regimes (arxiv.org)

I3L: Iterative Interactive Imitation Learning for Whole-Body Mobile Manipulation (arxiv.org)

Auditing Failure Attribution: Label Predictability and Judge Behavior (arxiv.org)

HiSTOR: Structured Treatment-outcome Representation Learning for High-Dimensional Longitudinal Data (arxiv.org)

Train for Many, Update with One: Improving Multi-Sample Reasoning via Fenchel-Bregman Posterior Learning (arxiv.org)

Explore with Memory, Learn without It: In-Context Exploration and Self-Distillation for LLM Reasoning (arxiv.org)

CoCA: Criterion-Aware Credit Assignment for Rubric-Based Reinforcement Learning (arxiv.org)

SkillScreen: High-Recall Screening of Malicious Agent Skills with Decision Models (arxiv.org)

TSRL: Trigger-based Supervised Reinforcement Learning for Continual Agent Training (arxiv.org)

Where Small Mistakes Matter: Quantifying State Criticality for Imitation Learning (arxiv.org)

From Token Correction to Claim Admission: Training-Free Hallucination Mitigation for Large Vision-Language Models (arxiv.org)

LeanPRM: Lean-Grounded Process Supervision for Natural Language Proof Verification (arxiv.org)

zPocket: Better Pocket Finding with Frozen Co-Folding Features (arxiv.org)

LatentRiskBench: Measuring State-Dependent Risks in Enterprise Agents (arxiv.org)

Dynamic K-Head: Difficulty-Supervised Budget Selection for Continuous Latent Reasoning (arxiv.org)

PrT-Flow: Probability-Tractable Flow Matching for Reinforcement Learning (arxiv.org)