Zhenghai Xue

SimpleTIR: End-to-End Reinforcement Learning for Multi-Turn Tool-Integrated Reasoning

Policy Regularization on Globally Accessible States in Cross-Dynamics Reinforcement Learning

Policy Optimization under Imperfect Human Interactions with Agent-Gated Shared Autonomy

Group-in-Group Policy Optimization for LLM Agent Training

AgentStudio: A Toolkit for Building General Virtual Agents

ASOR: Anchor State Oriented Regularization for Policy Optimization under Dynamics Shift