A Bit About Me

I am a researcher at Google DeepMind. Before joining industry, I earned my Ph.D. from the University of Utah. I live and work in Los Angeles.

The photo was taken at the Canyonlands National Park, Utah in December 2013.

• 6/5/25. After 15 months' wait, VideoPrism was finally open-sourced! Find it on GitHub and Hugging Face.
• 12/12/24. Our paper Video Creation by Demonstration was released. See results here.
• 5/1/24. Our paper VideoPrism: A Foundational Visual Encoder for Video Understanding was accepted to ICML 2024.
• 7/13/23. Our paper Unified Visual Relationship Detection with Vision and Language Models was accepted to ICCV 2023.
• 3/1/22. Our paper Contextualized Spatio-Temporal Contrastive Learning with Self-Supervision was accepted to CVPR 2022.
• 1/28/22. Our paper Surrogate Gap Minimization Improves Sharpness-Aware Training was accepted to ICLR 2022.
• 9/7/21. Our paper View-Invariant, Occlusion-Robust Probabilistic Embedding for Human Pose was accepted to International Journal of Computer Vision (IJCV). Code is released here.
• 3/3/21. Our paper Learning View-Disentangled Human Pose Representation by Contrastive Cross-View Mutual Information Maximization was accepted to CVPR 2021 with an oral presentation. Code is released here.
• 7/2/20. Our paper View-Invariant Probabilistic Embedding for Human Pose was accepted to ECCV 2020 with a Spotlight presentation. Code is released here.
• 10/27/19. Our DeepLab team won the 1st place in the Mapillary Vista Panoptic Segmentation Task at the ICCV 2019 COCO+Mapillary Vista Recognition Challenge, with paper Panoptic-DeepLab winning the Best Paper Award and the Most Innovative Award at the challenge workshop.
• 7/18/16. I joined the Mobile Vision Team at Google LAX.
• 7/11/16. Our paper SSHMT: Semi-Supervised Hierarchical Merge Tree for Electron Microscopy Image Segmentation was accepted to ECCV 2016 with a spotlight presentation. Code is released here.
• 7/5/16. Our paper Image Segmentation Using Hierarchical Merge Tree was accepted to IEEE Transactions on Image Processing. Code is released here.
• 4/29/16. Glia is released on Github.
• 8/24/15. I will be coordinating the imaging seminar this semester. Check out our website and feel free to present.​