Minghao Zhang


I am a senior undergraduate in the School of Software at Tsinghua University. During my undergraduate, I am lucky to work at UCSD (Xiaolong Wang's Lab), UC Berkeley (Koushil Sreenath's Lab ), and TuSimple. My research interest mainly lies in (deep) robotic learning, including reinforcement learning, legged locomotion, sim2real, and 3D vision.


Email: mehoozhang at gmail dot com

Github

Google Scholar

WeChat ID: mehooz (If you feel more comfortable contacting me in this way)

Papers

(* indicates equal contribution or alphabetical order)

Vision-Guided Quadrupedal Locomotion in the Wild with Multi-Modal Delay Randomization [webpage] [arxiv] [code]

Chiekio Sarah Imai*, Minghao Zhang*, Yuchen Zhang*, Marcin Kierebiński, Ruihan Yang, Yuzhe Qin, Xiaolong Wang (*ordered alphabetically)

(preprint)

Learning Vision-Guided Quadrupedal Locomotion End-to-End with Cross-Modal Transformers [webpage] [arxiv] [code]

Ruihan Yang*, Minghao Zhang*, Nicklas Hansen, Huazhe Xu, Xiaolong Wang

(ICLR 2022 spotlight)

DAIR: Disentangled Attention as Intrinsic Regularization for Collaborative Bimanual Manipulation [webpage] [arxiv]

Minghao Zhang*, Pingcheng Jian*, Yi Wu, Huazhe Xu, Xiaolong Wang

(ICML 2021 workshop)

Bridging Imagination and Reality for Model-Based Deep Reinforcement Learning [arxiv] [code]

Minghao Zhang*, Guangxiang Zhu*, Honglak Lee, Chongjie Zhang

(NeurIPS 2020)

View-Aware Joint Structure-Geometry Learning for Single-View 3D Shape Reconstruction [paper]

Xuancheng Zhang*, Rui Ma*, Changqing Zou, Minghao Zhang, Xibin Zhao, Yue Gao

(TPAMI 2021)

Projects

Tianshou: An elegant, flexible, and superfast PyTorch deep Reinforcement Learning platform [code] [arxiv]

(An open-source RL platform)

Services

Reviewer for RA-L, ICML 2022, NeurIPS 2022