Yao (Mark) Mu

Bio: I am currently a Ph.D. Student of Computer Science at the University of Hong Kong, fortunately supervised by Prof. Ping Luo and co-advised by Prof. Xiaoou Tang at the MMLab@HKU. Previously I obtained the M.Phil Degree from Tsinghua University in June 2021,under the supervision of Prof. Bo Cheng and Prof. Shengbo Li at the Intelligent Driving Laboratory and Prof. Jianyu Chen at the Institute for Interdisciplinary Information Sciences in Tsinghua University. I work closely with Dr. Mingyu Ding at UC Berkeley.
Contact me with email: muyao@connect.hku.hk and Zhihu: 知乎.
Research Interests: Embodied AI, Generative Robot Agent, Reinforcement Learning, Roboitcs Control. My goal is to build a general embodied AI system that can learn optimal policy efficiently and can be generalized well in unseen tasks and environments. My research slogan: Shape the intelligence, Spirit the machine! (为智慧塑形，为机器注灵！)
Joining My Group: I am looking for student researchers and visiting students to join my group. Please feel free to send me an email with your resume if you are interested in doing research with me. Enjoy life, enjoy research. May the force be with you.
News: I am the primary responsible person for DeepTimber Robotics Innovations Community(深木社区->向上开花，向下扎根): A Community for Multi-modal Embodied AI and Robotics Enthusiasts, a large community of engineers and researchers who share interests in embodied AI and robotics. Please find some details about DeepTimber Robotics Innovations Community in this link and this github page.
News: Our paper "Skilldiffuser: Interpretable hierarchical planning via skill abstractions in diffusion-based task execution" is accepted by CVPR 2024!
News: Three Papers[SPET, Tree-Planner, Aligndiff] are accepted by ICLR 2024!
News: Our paper "EmbodiedGPT: Vision-Language Pre-Training via Embodied Chain of Thought" has been accepted by NeurIPS 2023 (Selected as Spotlight)!
News: Our paper "AdaptDiffuser: Diffusion Models as Adaptive Self-evolving Planners" has been accepted by ICML 2023 (Oral Presentation)!
News: Our paper "MetaDiffuser: Diffusion Model as Conditional Planner for Offline Meta-RL." has been accepted by ICML 2023!
News: Our paper "EC^2: Emergent Communication for Embodied Control." has been accepted by CVPR 2023!
News: Our paper "CO^3: Cooperative Unsupervised 3D Representation Learning for Autonomous Driving" has been accepted by ICLR 2023!
News: Our paper "EUCLID: Towards Efficient Unsupervised Reinforcement Learning with Multi-choice Dynamics Model" has been accepted by ICLR 2023!
2022-09:Our paper "DOMINO: Decomposed Mutual Information Optimization for Generalized Context in Meta-Reinforcement Learning" has been accepted by NeurIPS 2022(Spotlight)!
2022-03: Our paper "CtrlFormer: Learning Transferable State Representation for Visual Control via Transformer" has been accepted by ICML 2022!
Awards:
Hong Kong PhD Fellowship Scheme (HKPFS)
HKU Presidential PhD Scholar Programme (HKU-PS)
Student Best Paper Award at the 20th ICCAS
Finalists for the Best Student Paper Award of IV2021 (3/450)
Outstanding Graduate of Tsinghua University (only 1 person/faculty)
Outstanding Thesis Award, Tsinghua University
National Scholarship (3 times)

Talks

---

Talk at RLChina:

AdaptDiffuser: Diffusion Models as Adaptive Self-evolving Planners

Link

---

Talk at New Youth:

Knowledge Transfer and Generalization of Deep Reinforcement Learning

Link