Our laboratory focuses on Egocentric Vision and Multimodal Generative AI. We are committed to:

- Designing AI that sees through your eyes, learns your skills, and understands your intentions. (Developing the next generation of human-centric intelligent systems that can see what you see, learn what you know, and understand what you think.)

I am currently a Senior Research Scientist at Meta GenAI. I received my Ph.D. from Georgia Institute of Technology and have been selected as a Young Talent for the National High-Level Talent Program (Overseas). I am dedicated to the theoretical research of egocentric vision and multimodal generative AI. I have published over 20 papers in top conferences and journals such as CVPR, ECCV, ACL, and TPAMI as a primary author. My research has been cited more than 9,000 times, with eight papers selected for oral presentations. My papers at CVPR 2022 and ECCV 2024 were shortlisted for the Best Paper Award, and the paper at BMVC 2022 received the Best Student Paper Award.

As a primary author, I have contributed to the construction of several widely recognized egocentric video datasets, including Ego4D, EgoExo4D, EGTEA Gaze+, and Behavior Vision Suite. These datasets have been widely applied in both academia and industry. In the field of first-person behavior understanding, I have proposed several algorithms for action recognition and prediction, which will be applied to the next generation of smart glasses products at Meta Reality Labs. Additionally, during my tenure at Meta GenAI, I have been deeply involved in the training and evaluation of multimodal generative models, including EMU, Llama3, and Llama4 (the multimodal parts only).