强化学习：
- Offline Reinforecement Learning
- Diffusion/Flow based Reinforecement Learning
- Hierarchical Reinforecement Learning
模仿学习：
- Manipulation for Robot Arm and Dexterous Hand
- Vision-Language-Action/Navigation (VLA/VLN)
- Real-Time Action Chunking (RTC)，并行推理
- VLA Distillation for Embodied System Acceleration
多智能体系统：
- Multi-agent Theory for LLM
- Homogeneous/Heterogeneous Agent Collaboration
- Multimodal Collaborative Perception
以上方向本人均可招收研究生并提供较为深入的科研指导，对纯理论方向感兴趣的同学可偏好强化学习或Multi-agent Theory for LLM等领域，模仿学习方向更注重工程实战，非常欢迎对3D视觉感兴趣的同学私信合作！