Greetings from Berkeley~
The Berkeley Artificial Intelligence Research (BAIR) seminar series is hosted by the BAIR Lab and showcases dissertations and job talks of BAIR Ph.D. students and postdoctoral researchers.
These talks are held in person on the 8th floor of Berkeley Way West (BWW) at 11:10 am on Wednesdays. For those who join in person, the seminar will be followed by pizza in the kitchen of BWW8. If you are unable to attend in person, please use this link to join seminar over Zoom. (Password: 8006)
Wednesday, 5/7's speaker is Jathushan Rajasegaran, a Ph.D. student advised by Prof. Jitendra Malik. Jathushan will present his dissertation.
Title: Video Models of People and Pixels
Abstract: From the moment we are born, we continuously witness the “video” of our own lives—hundreds of thousands of hours of rich, unfolding scenes. In this talk, I will present approaches to build video models to understand the world and humans, including tracking, recognition and prediction. I'll begin with how to represent people in 3D space, and how we used these 3D features to track people in internet style videos. Next, once we have a good tracking system in 3D, we can automatically scale up tracking data and learn a model that can reason about human actions. Then I'll talk about training large scale video prediction models directly from pixels, and showcase their versatility—from recognition and tracking to object permanence and robotics. Finally, I will discuss scaling behaviours of visual models, from autoregressive and diffusion to clip models.
Bio: Jathushan Rajasegaran is a Ph.D student in the Department of EECS at UC Berkeley and a member of BAIR, advised by Prof. Jitendra Malik. He obtained his B.Sc. degree in Electronic and Telecommunications at University of Moratuwa, Sri Lanka. He is also a visiting researcher at FAIR. His research interests lie in the general area of computer vision and deep learning, particularly in learning from videos, self-supervised learning and deep neural architectures.
Wednesday seminar links: (Please do not distribute the below links beyond your organization.)
For those of you who are unable to join in person, please join via Zoom here (Password: 8006)
Unpublished Playlist of Talk Recordings
Unpublished Playlist of 2024 Recordings
Speaker Slides
Click here to add the BAIR Seminar to your calendar
Hope to see you at a seminar~
The BAIR Admin Team
The Berkeley Artificial Intelligence Research (BAIR) seminar series is hosted by the BAIR Lab and showcases dissertations and job talks of BAIR Ph.D. students and postdoctoral researchers.
These talks are held in person on the 8th floor of Berkeley Way West (BWW) at 11:10 am on Wednesdays. For those who join in person, the seminar will be followed by pizza in the kitchen of BWW8. If you are unable to attend in person, please use this link to join seminar over Zoom. (Password: 8006)
Wednesday, 5/7's speaker is Jathushan Rajasegaran, a Ph.D. student advised by Prof. Jitendra Malik. Jathushan will present his dissertation.
Title: Video Models of People and Pixels
Abstract: From the moment we are born, we continuously witness the “video” of our own lives—hundreds of thousands of hours of rich, unfolding scenes. In this talk, I will present approaches to build video models to understand the world and humans, including tracking, recognition and prediction. I'll begin with how to represent people in 3D space, and how we used these 3D features to track people in internet style videos. Next, once we have a good tracking system in 3D, we can automatically scale up tracking data and learn a model that can reason about human actions. Then I'll talk about training large scale video prediction models directly from pixels, and showcase their versatility—from recognition and tracking to object permanence and robotics. Finally, I will discuss scaling behaviours of visual models, from autoregressive and diffusion to clip models.
Bio: Jathushan Rajasegaran is a Ph.D student in the Department of EECS at UC Berkeley and a member of BAIR, advised by Prof. Jitendra Malik. He obtained his B.Sc. degree in Electronic and Telecommunications at University of Moratuwa, Sri Lanka. He is also a visiting researcher at FAIR. His research interests lie in the general area of computer vision and deep learning, particularly in learning from videos, self-supervised learning and deep neural architectures.
Wednesday seminar links: (Please do not distribute the below links beyond your organization.)
For those of you who are unable to join in person, please join via Zoom here (Password: 8006)
Unpublished Playlist of Talk Recordings
Unpublished Playlist of 2024 Recordings
Speaker Slides
Click here to add the BAIR Seminar to your calendar
Hope to see you at a seminar~
The BAIR Admin Team
Berkeley Artificial Intelligence Research (BAIR) Lab