Zeshi Yang (楊 澤世)

I am an eclectic independent researcher, focusing on developing physically grounded and emotionally rich humanoid agents. I have a lot of experience at realizing agile and dexterous motions on simulated humanoids. I received my Ph.D. in 2023 from Simon Fraser University, where I worked on character animation, motion control and deep reinforcement learning. Prior to that, I earned a bachelor's degree in applied physics from USTC in 2018.
Email Google Scholar Zhihu

Education & Career

Education

USTC
B.S. in Applied Physics
University of Science and Technology of China
2014 - 2018
SFU
Ph.D in Computer Science
Simon Fraser University
2018 - 2023

Career

PKU
Visiting Scholar
Peking University, China
2021 Aug - 2022 Aug
LightSpeed
Research Intern
LightSpeed Studios, Tencent America
2022 Sep - 2023 May
miHoYo
Research Scientist
miHoYo
2023 June - 2024 Sep
Anuttacon
Research Scientist
Anuttacon
2024 Oct - Now

Research Interests

Video Generation as World Models World Action Models Motion Synthesis & Control Reinforcement Learning

Industry projects

LPM 1.0 (Large Performance Model) | 2026
Infinite-length, realtime, realistic streaming video generation model.
Tech lead of the online video generation model.
Project Page Technical Report

Whispers from the Star | 2025
An AI-driven game with emotionally rich humanoid agents.
Tech lead of the motion generation system.
Steam YouTube Bilibili

Publications

TokenHSI: Unified Synthesis of Physical Human-Scene Interactions through Task Tokenization
Liang Pan, Zeshi Yang, Zhiyang Dou, Wenjia Wang, Buzhen Huang, Bo Dai,
Taku Komura, Jingbo Wang†
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2025)
Paper
Real-time Diverse Motion In-betweening with Space-time Control
Yuchen Chu, Zeshi Yang
ACM SIGGRAPH Conference on Motion, Interaction and Games (MIG 2024)
Learning-based 2D Irregular Shape Packing
Zeshi Yang, Zherong Pan, Manyi Li, Kui Wu, Xifeng Gao
ACM Transactions on Graphics, Volume 42, Article 6 (Proc. ACM SIGGRAPH Aisa 2023)
Acquiring Stylized Motor Skills for Physics-based Characters
PhD Thesis, Simon Fraser University, 2023
Learning to Use Chopsticks in Diverse Gripping Styles
Zeshi Yang, KangKang Yin, Libin Liu
ACM Transactions on Graphics, Volume 41, Issue 4, Article 95 (Proc. ACM SIGGRAPH 2022)
Efficient Hyperparameter Optimization for Physics-based Character Animation
Zeshi Yang, Zhiqi Yin
ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games 2021 (I3D 2021)
Discovering Diverse Athletic Jumping Strategies
Zhiqi Yin, Zeshi Yang, Michiel van de Panne, KangKang Yin
ACM Transactions on Graphics, Volume 40, Issue 4, Article 91 (Proc. ACM SIGGRAPH 2021)
Neural fidelity warping for efficient robot morphology design
Sha Hu, Zeshi Yang, Greg Mori
2021 International Conference on Robotics and Automation (ICRA 2021)
Learning and Exploring Motor Skills with Spacetime Bounds
Li-Ke Ma, Zeshi Yang, Xin Tong, Baining Guo, KangKang Yin
Computer Graphics Forum (Eurographics), 40, 2(2021)
Improving Skeleton-based Action Recognition with Robust Spatial and Temporal Features
Zeshi Yang, Kangkang Yin
In Arxiv 2020
Towards Robust Direction Invariance in Character Animation
Li-Ke Ma, Zeshi Yang, Baining Guo, KangKang Yin
Computer Graphics Forum (Pacific Graphics), 39, 7(2019)
Redirected Smooth Mappings for Multiuser Real Walking in Virtual Reality
Zhi-Chao Dong, Xiao-Ming Fu, Zeshi Yang, Ligang Liu
ACM Transactions on Graphics, 38(5), 2019.(Presented at Siggraph Asia)

Contact

Email: zs243[at]mail.ustc.edu.cn
Email: zeshiyang209[at]gmail.com
© 2026 Zeshi Yang. Last updated April 2026.