Ce Hao has graduated and is now Dr. Hao. Congratulations Ce!
Ce’s PhD thesis, Advanced Diffusion Robot Manipulation Policies for Language Reasoning, Long-Horizon Planning, and Skill Abstraction, advanced diffusion models for robot learning across three themes:
- Language reasoning. DISCO grounds language-guided manipulation in diffusion policies via constrained inpainting, and SkillVLA brings skill reuse to vision–language–action models for dual-arm manipulation.
- Long-horizon planning. CHD couples hierarchical diffusion across temporal scales to solve long-horizon tasks.
- Skill abstraction. SMP learns a compact set of reusable mixture-of-experts skill experts, and Guided-SSIP adds inference-time guidance to streaming diffusion policies.
He also explored tactile-informed policies with VLA-Touch. We’re proud of everything Ce accomplished and can’t wait to see what’s next!
