Tongzhou Mu 木同舟

Tongzhou Mu 

Ph.D. candidate,
Department of Computer Science and Engineering,
University of California San Diego
Email: t3mu [@] ucsd [DOT] edu
Google Scholar / DBLP / Twitter / Github / LinkedIn

I am currently working on learning from demonstrations. Email me if you are interested in collaboration!

About me

I am a Ph.D. candidate in the Department of Computer Science and Engineering at the University of California San Diego, and I am fortunate to be advised by Prof. Hao Su. My long-term research goal is to build a decision-making framework with strong generalizability. Specifically, I am interested in Reinforcement Learning / Imitation Learning, Concept Discovery and Reasoning, and Robotics / Embodied AI.

Prior to starting as a Ph.D. student, I received my B.Eng. and M.Sc. degrees in computer science from Zhejiang University, and UC San Diego in 2017 and 2019, respectively. I have also spent time at Microsoft Research Asia, Intel AI, Wormpex AI , Amazon Alexa AI, and NVIDIA Robotics Lab.

I am the major developer of the 1st version of ManiSkill benchmark, and the lead organzier of the ManiSkill Challenge 2021 and the Generalizable Policy Learning in the Physical World Workshop. I also actively contribute to the development of the subsequent generations of ManiSkill, e.g., ManiSkill 2 and ManiSkill 3.

Research interests

My research interests include

  • Reinforcement Learning / Imitation Learning

    • Sturctured Policies and Neural Programs

    • Learning from Demonstrations

    • Curriculum in RL

  • Concept Discovery

    • Emerged Concepts from Interactions

    • Reasoning over Concepts

  • Robotics / Embodied AI

    • Planning and Control for Manipulation

    • Building Full-Physical Tasks in Simulator

Publications & Preprints

Papers sorted by years. Representative papers are highlighted.



Learning Reusable Dense Rewards for Multi-Stage Tasks
Tongzhou Mu, Minghua Liu, Hao Su
Submitted to ICLR 2024
[Project Page]


On Pre-Training for Visuo-Motor Control: Revisiting a Learning-from-Scratch Baseline
Nicklas Hansen*, Zhecheng Yuan*, Yanjie Ze*, Tongzhou Mu* , Aravind Rajeswaran+, Hao Su+, Huazhe Xu+, Xiaolong Wang+
International Conference on Machine Learning (ICML) 2023
Pre-training Robot Learning Workshop at CoRL 2022


Abstract-to-Executable Trajectory Translation for One-Shot Task Generalization
Stone Tao, Xiaochen Li, Tongzhou Mu, Zhiao Huang, Yuzhe Qin, Hao Su
International Conference on Machine Learning (ICML) 2023
Deep Rieinforcement Learning Workshop at NeurIPS 2022
[Project Page] [Video] [Code] [Slides]
* Stone Tao and Xiaochen Li are undergruduate students I advised


Boosting Reinforcement Learning and Planning with Demonstrations: A Survey
Tongzhou Mu, Hao Su
ArXiv preprint


ManiSkill2: A Unified Benchmark for Generalizable Manipulation Skills
Jiayuan Gu, Fanbo Xiang, Xuanlin Li*, Zhan Ling*, Xiqiang Liu*, Tongzhou Mu* , Yihe Tang*, Stone Tao*, Xinyue Wei*, Yunchao Yao*, Xiaodi Yuan, Pengwei Xie, Zhiao Huang, Rui Chen, Hao Su
* equally contributed authors are ordered by alphabets
International Conference on Learning Representations (ICLR) 2023
[Code] [Project Page] [Challenge Website]


Close the Optical Sensing Domain Gap by Physics-Grounded Active Stereo Sensor Simulation
Xiaoshuai Zhang, Rui Chen, Fanbo Xiang, Yuzhe Qin, Jiayuan Gu, Zhan Ling, Minghua Liu, Peiyu Zeng, Songfang Han, Zhiao Huang, Tongzhou Mu, Jing Xu, Hao Su
IEEE Transactions on Robotics (T-RO) 2023



Learning Two-Step Hybrid Policy for Graph-Based Interpretable Reinforcement Learning
Tongzhou Mu, Kaixiang Lin, Feiyang Niu, Govind Thattai
Transactions on Machine Learning Research (TMLR) 2022
Elements of Reasoning: Objects, Structure, and Causality Workshop at ICLR 2022



ManiSkill: Generalizable Manipulation Skill Benchmark with Large-Scale Demonstrations
Tongzhou Mu* , Zhan Ling*, Fanbo Xiang*, Derek Yang*, Xuanlin Li*, Stone Tao, Zhiao Huang, Zhiwei Jia, Hao Su
Conference on Neural Information Processing Systems (NeurIPS) Datasets and Benchmarks Track 2021
[Video] [Slides] [Poster] [Code] [Baselines] [Challenge Website]



Refactoring Policy for Compositional Generalizability using Self-Supervised Object Proposals
Tongzhou Mu* , Jiayuan Gu*, Zhiwei Jia, Hao Tang, Hao Su
Conference on Neural Information Processing Systems (NeurIPS) 2020
[Project Page] [Poster] [Slides] [Code]


State Alignment-based Imitation Learning
Fangchen Liu, Zhan Ling, Tongzhou Mu, Hao Su
International Conference on Learning Representations (ICLR) 2020

Before 2019


Transfer Value or Policy? A Value-centric Framework Towards Transferrable Continuous Reinforcement Learning
Xingchao Liu*, Tongzhou Mu* , Hao Su
Deep Reinforcement Learning Workshop at NeurIPS 2018


Accelerated Doubly Stochastic Gradient Algorithm for Large-scale Empirical Risk Minimization
Zebang Shen, Hui Qian, Tongzhou Mu, and Chao Zhang
International Joint Conference on Artificial Intelligence (IJCAI) 2017


Adaptive Variance Reducing for Stochastic Gradient Descent
Zebang Shen, Hui Qian, Tengfei Zhou, and Tongzhou Mu
International Joint Conference on Artificial Intelligence (IJCAI) 2016


Professional Services


  • Teaching Assistant: CSE 152 Introduction to Computer Vision at UC San Diego, Fall 2018

  • Consultant Volunteer: CSE291-J00 Deep Learning Lab (Computer Vision) at UC San Diego, Fall 2020

  • Guest Lecturer: CSE291-A00 Machine Learning for Robotics at UC San Diego, Winter 2023