Stanford CS521 - AI Safety Seminar

Lecture Details

Zico Kolter, Associate Professor, Carnegie Mellon University
May 4, 2022

This talk looks at the task of verifying deep networks, guaranteeing that outputs of a network obey certain properties for certain classes of inputs. Such approaches can be used to validate robustness and safety of neural networks, but such exact verification is a hard combinatorial problem, and off-the-shelf solvers typically perform quite poorly. Nonetheless, over the past several years there has been a large amount of progress in the area, and recent methods are able to verify medium-sized networks thousands of times faster than generic solvers. In this talk, I will provide a general overview of the verification problem, then highlight several advances we and others have made in recent years to achieve these speedups. Many of these approaches were implemented in our submission to the Verification of Neural Networks Competition (VNNCOMP) in 2021, where our team, a collaboration with UCLA, Northeastern, and Columbia, won first place across most categories.

About the speaker: Zico Kolter is an Associate Professor in the Computer Science Department with the School of Computer Science at Carnegie Mellon University. In addition to his full-time role at CMU, he also serves as Chief Scientist of AI Research for the Bosch Center for AI (BCAI), working in the Pittsburgh Office.