Lei Han

3 papers Tencent AI Lab

• Learning 3D Persistent Embodied World Models
• Multi-Agent Collaboration via Evolving Orchestration
• STAR: Efficient Preference-based Reinforcement Learning via Dual Regularization