Xiaoyu Tan

PRISM: Festina Lente Proactivity—Risk-Sensitive, Uncertainty-Aware Deliberation for Proactive Agents

The Choice of Divergence: A Neglected Key to Mitigating Diversity Collapse in Reinforcement Learning with Verifiable Reward

Count Counts: Motivating Exploration in LLM Reasoning with Count-based Intrinsic Rewards

CUARewardBench: Benchmark for Evaluating Reward Models on Computer-using Agent Trajectories

Learn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning

Asleep at the Wheel: Benchmarking the Inattention of Vision Language Models to Clinical Sleep Signals

Distributional Reinforcement Learning for Large Language Models

Training-Free Group Relative Policy Optimization

Transitioning from Full-Context to Active Evidence-Seeking Evaluation: A Novel Benchmark for Real-World Artificial Intelligence Assisted Medical Diagnosis

Dual-Route Mental Imagery for Robust VLM-based Medical Image Diagnosis

RoRecomp: Enhancing Reasoning Efficiency via Rollout Response Recomposition in Reinforcement Learning

MMUDA: Towards Robust Sleep Staging with Multi-Source Multi-Channel Unsupervised Domain Adaptation

ORIGAMISPACE: Benchmarking Multimodal LLMs in Multi-Step Spatial Reasoning with Mathematical Constraints

Atomic Thinking of LLMs: Decoupling and Exploring Mathematical Reasoning Abilities

Refine Knowledge of Large Language Models via Adaptive Contrastive Learning

One Example Shown, Many Concepts Known! Counterexample-Driven Conceptual Reasoning in Mathematical LLMs