Highlighted research talks

The Platonic Representation Hypothesis
Simons Institute, June 2024.
Generative Models as Data++
Generative Models for Computer Vision Workshop at CVPR, June 2023.
Structures and Environments for Generalist Agents
CMU Robotics Insitute Seminar, April 2023.
When and Why Does Contrastive Learning Work?
L2ID Workshop at CVPR, June 2021.

Additional research talks

June 2024, LIMIT Workshop at CVPR: N=0: Learning Vision Without Visual Data
December 2022, Robotics Seminar, MIT: Giving Robots Mental Imagery
Nov 2022, BMVC Keynote When faking your data actually helps -- Learning vision from GANs, NeRFs, and Noise
June 2022, Artificial Social Intelligence Workshop at CVPR: Three Lessons from Computer Vision, for Social Intelligence
June 2022, MultiEarth Workshop at CVPR: Hallucinating Multiview Data for Computer Vision
October 2021, Samsung AI Forum: Data++: Exploring Model-Based Data for AI
Feb 2020, Embodied Intelligence Seminar, MIT: Emergent Intelligence: Getting more out of your agents than you bake in
Feb 2017, University of Washington: Learning to See Without a Teacher

Educational Talks and Tutorials

January 2024, Expanding Horizons in Computing (at MIT SCC): Deep Learning Bootcamp