Yue Deng

Value-aligned World Model Regularization for Model-based Reinforcement Learning

RF-Agent: Automated Reward Function Design via Language Agent Tree Search

Physics-aligned field reconstruction with diffusion bridge

Value-aligned Behavior Cloning for Offline Reinforcement Learning via Bi-level Optimization

Progress Reward Model for Reinforcement Learning via Large Language Models

Hybrid Regularization Improves Diffusion-based Inverse Problem Solving

COSTAR: Dynamic Safety Constraints Adaptation in Safe Reinforcement Learning