Jiafeng Xu

Jiafeng Xu

Robotics Researcher, ByteDance Seed Robotics
Previously Researcher, Tencent Robotics X
Robot manipulation·Legged locomotion·Reinforcement learning·Optimal control

About Me

I am currently a robotics researcher at ByteDance Seed-Robotics, where my research focuses on robotic reinforcement learning, learning-based robot manipulation and locomotion, and unified control frameworks for robots. Previously, I worked as a researcher at Tencent Robotics X, with research interests in robot motion control and planning, high-dynamic motion planning, and real-time dynamic parameter identification for robotic systems. I hold both a bachelor's and a master's degree from the Beijing Advanced Innovation Center for Intelligent Robots and Systems at Beijing Institute of Technology, where I was supervised by Professors Zhihong Jiang and Qiang Huang.

I am deeply passionate about robotics, with my career consistently centered on robot control and planning. During my tenure at Tencent, I concentrated on real-time control and planning grounded in dynamics and numerical optimization. At ByteDance, I have conducted in-depth research on end-to-end control methods for large-scale neural networks trained using supervised learning and reinforcement learning. My long-term ambition is to develop general-purpose embodied robots with human-like intelligence, enabling robots to seamlessly integrate into human production and daily life, and ultimately drive meaningful progress in human society.

Research Interests

Contact me: if you are interested in collaborating or discussing any research question, feel free to drop me an email at chnjiafengxu@gmail.com.

News

Publications

For a full and up-to-date list, please see my Google Scholar. (*: equal contribution)

Technical Reports

GR-RL
GR-RL: Going Dexterous and Precise for Long-Horizon Robotic Manipulation
Technical Report, ByteDance Seed Robotics, 2025
[PDF] [arXiv] [Website]
ByteWrist
ByteWrist: A Parallel Robotic Wrist Enabling Flexible and Anthropomorphic Motion for Confined Spaces
Technical Report, ByteDance Seed Robotics, 2025
[PDF] [arXiv] [Website] [Seed]
Manipulation as in Simulation
Manipulation as in Simulation: Enabling Accurate Geometry Perception in Robots
Technical Report, ByteDance Seed Robotics, 2025
[PDF] [arXiv] [Website]
GR-3
GR-3 Technical Report
Technical Report, ByteDance Seed Robotics, 2025
[PDF] [arXiv] [Website]
GR-2
GR-2: A Generative Video-Language-Action Model with Web-Scale Knowledge for Robot Manipulation
Technical Report, ByteDance AI Lab Robotics Research, 2024
[PDF] [arXiv] [Website]

Conference & Journal Papers

Flow-Based Policy for Online RL
Flow-Based Policy for Online Reinforcement Learning
Lei Lv, Yunfei Li, Yu Luo, Fuchun Sun, Tao Kong, Jiafeng Xu*, Xiao Ma
Conference on Neural Information Processing Systems (NeurIPS), 2025
[PDF] [arXiv] [Code]
WMP
World Model-based Perception for Visual Legged Locomotion
Hang Lai, Jiahang Cao, Jiafeng Xu*, Hongtao Wu, Yunfeng Lin, Tao Kong, Yong Yu, Weinan Zhang
IEEE International Conference on Robotics and Automation (ICRA), 2024
[PDF] [arXiv] [Code] [Website]
GR-1
GR-1: Unleashing Large-Scale Video Generative Pre-training for Visual Robot Manipulation
Hongtao Wu, Ya Jing, Chilam Cheang, Guangzeng Chen, Jiafeng Xu, Xinghang Li, Minghuan Liu, Hang Li, Tao Kong
International Conference on Learning Representations (ICLR), 2023
[PDF] [arXiv] [Code] [Website]
MOMA-Force
MOMA-Force: Visual-Force Imitation for Real-World Mobile Manipulation
Taozheng Yang, Ya Jing, Hongtao Wu, Jiafeng Xu*, Kuankuan Sima, Guangzeng Chen, Qie Sima, Tao Kong
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2023
[PDF] [DOI] [Website]
Locomotion Generation for Quadruped Robots
Locomotion Generation for Quadruped Robots on Challenging Terrains via Quadratic Programming
Xinyang Jiang, Wanchao Chi, Yu Zheng, Shenghao Zhang, Yonggen Ling, Jiafeng Xu, Zhengyou Zhang
Autonomous Robots, 2022
[PDF] [DOI]
Real-time Inertial Parameter Identification
Real-time Inertial Parameter Identification of Floating-Base Robots Through Iterative Primitive Shape Division
Jiafeng Xu, Yu Zheng, Xinyang Jiang, Sicheng Yang, Lingzhu Xiang, Zhengyou Zhang
IEEE International Conference on Robotics and Automation (ICRA), 2022
[PDF] [DOI]
Stable Parking Control of a Robot Astronaut
Stable Parking Control of a Robot Astronaut in a Space Station Based on Human Dynamics
Zhihong Jiang, Jiafeng Xu, Hui Li, Qiang Huang
IEEE Transactions on Robotics (T-RO), 2019
[PDF] [DOI]
ICARM
2019
The Design and Development of an Anthropomorphic Worm-Gear Driven Robotic Hand: BIT-JOCKO
Changjin Wang, Yao Sun, Jiafeng Xu, Xingdong Liu, Xinbing Zhou, Xiaopeng Chen
IEEE International Conference on Advanced Robotics and Mechatronics (ICARM), 2019
[DOI]
Best Paper Finalist
IEEE ISR
2018
Trajectory Planning Based on Dual Torque Feedforward Control for Robot Astronaut Safe Movement on Space Station
Zhihong Jiang, Jiafeng Xu, Hui Li, Qiang Huang
IEEE International Conference on Intelligence and Safety for Robotics (ISR), 2018
[DOI]
Best Paper Finalist

Projects

GR-RL
GR-RL
Dexterous and precise long-horizon manipulation with RL.
Manipulation as in Simulation
Manipulation as in Simulation
Accurate geometry perception for real-robot manipulation.
GR-3
GR-3
A large-scale vision-language-action model for robots.
GR-2
GR-2
Generative video-language-action model with web-scale knowledge.
WMP
WMP
World model-based perception for visual legged locomotion.
GR-1
GR-1
Video generative pre-training for visual robot manipulation.
MOMA-Force
MOMA-Force
Visual-force imitation for real-world mobile manipulation.
Max
Max
High-dynamic quadruped robot (Tencent Robotics X).
Ollie
Ollie
Whole-body motion control of a wheel-legged robot (Tencent Robotics X).

Awards & Honors