معرفی
Xue Bin Peng is an Assistant Professor at Simon Fraser University and a Research Scientist at NVIDIA. He holds a Ph.D. from UC Berkeley (advised by Sergey Levine and Pieter Abbeel) and an M.Sc. from the University of British Columbia (advised by Michiel van de Panne). His research focuses on the intersection of computer graphics and machine learning, particularly reinforcement learning for motion control of simulated characters. Prior roles include internships at Sony, Google Brain, OpenAI, Adobe Research, Disney Research, Microsoft, and Capcom.
Key research areas include physically-based character animation, motion imitation, and sim-to-real robotic control. His work emphasizes scalable methods for generating diverse, adaptive, and realistic motions, often leveraging diffusion models, adversarial learning, and multi-task frameworks. Notable contributions include DeepMimic, AMP, and CLoSD, which address challenges in motion control and style transfer. He has received awards such as the Best Paper Honourable Mention (SIGGRAPH 2023) and Best Paper Award (RSS 2020).
His publications span top venues like ACM SIGGRAPH, ICLR, NeurIPS, and ICRA, with a focus on advancing reinforcement learning techniques for robotics and animation. Collaborations include NVIDIA, UBC, and institutions like Max Planck Institute and Carnegie Mellon University.
