|
|
XuDong Wang (王旭东)How to pronounce XuDong? It is composed of two syllables: "Xu" and "Dong": /ʃü/-/dɔ:ŋ/Hi! I am a Member of Technical Staff at Physical Intelligence, where I lead the Video-Action-Models team, which aims to build generalizable robot intelligence through world modeling, and an incoming Assistant Professor at Duke University. I received my Ph.D. in EECS from UC Berkeley, where I was a member of the Berkeley AI Research (BAIR) Lab and was advised by Prof. Trevor Darrell. I also received my M.S. (thesis track) in Intelligent Systems, Robotics, and Control from UC San Diego, where I was advised by Prof. Nuno Vasconcelos. Before that, I earned my B.S. from Jilin University as an honorary member of the Tang Aoqing Honors Program in Science.I was a Research Scientist at Meta Superintelligence Labs (Dec 2024-Dec 2025), where I worked on native video generation models. Before that, I was a student researcher at Google DeepMind (Mountain View, 2024-2025), working with Prof. Cordelia Schmid, Dr. Xingyi Zhou, and Dr. Alireza Fathi. Before that, I interned at Meta AI—with the GenAI team (Seattle, 2023) and FAIR (New York, 2022)—hosted by Dr. Ishan Misra and Dr. Rohit Girdhar. Before joining BAIR, I was a member of the International Computer Science Institute (ICSI), working closely with Prof. Stella X. Yu, Dr. Zhirong Wu, and Prof. Ziwei Liu. My research has been primarily focused on self-supervised learning, multimodality models and machine learning. I am particularly interested in advancing multimodality models beyond human supervision. By enhancing self-supervised learning techniques for multimodality models—minimizing reliance on human-annotated datasets—I strive to develop intelligent systems that can comprehend and interact with the environment in ways that mirror or even surpass human sensation, perception, and reaction. Contact: xdwang [at] eecs [dot] berkeley [dot] edu |
