AIR Distinguished Speaker Series: Deva Ramanan, Professor, Carnegie Mellon University

Talk title: Perceiving and Understanding a Dynamic 3D World

Talk abstract: Massively-trained data-driven models have produced remarkable progress in computer vision and natural language processing. In this talk, I will review various methods for extending such approaches to the physical world. The central thesis is that one will need to learn structured representations from streaming sensor data, beit from everyday monocular video footage or lidar logs from autonomous driving fleets. However, rather than simply predicting the next sensor frame of pixels or depth measurements, we predict compositional 4D representations of a dynamic world, with a particular emphasis on the prediction of deformable objects such as people and animals. To do so, we combine data-driven learning with an analysis-by-synthesis framework, combining recent innovations in generative models with differentiable rendering and differentiable simulation. I will present initial experiments that suggest such representations can be used for a variety of downstream applications such as large-scale reconstructions, augmented reality, and robotic navigation in dynamic environments.

Bio: Deva Ramanan is a Professor in the Robotics Institute at Carnegie-Mellon University and the former director of the CMU Center for Autonomous Vehicle Research. His research interests span computer vision and machine learning, with a focus on visual understanding of videos. He has been awarded the IEEE PAMI Young Researcher Award, named one of Popular Science's Brilliant 10 researchers, named a National Academy of Sciences Kavli Fellow, won the Longuet-Higgins Prize for fundamental contributions in computer vision, and was recognized for best paper awards in all major vision conferences (CVPR, ECCV, ICCV). His work has been supported by NSF, ONR, DARPA, as well as industrial collaborations with Intel, Google, Microsoft, and Meta.

When 11:00 am to 12:00 pm on Wednesday, April 10, 2024
Building CDS, 665 Commonwealth Ave
Room Room 1101 (11th floor)
Fees Free
Speakers Deva Ramanan, Professor of Robotics at Carnegie Mellon University