A Picture of the Energy Landscape of Deep Neural Networks

Monday, October 29, 2018 - 1:25pm - 2:25pm
Lind 305
Pratik Chaudhari (California Institute of Technology)
Deep networks are mysterious. These over-parametrized machine learning models, trained with rudimentary optimization algorithms on non-convex landscapes in millions of dimensions have defied attempts to put a sound theoretical footing beneath their impressive performance.

This talk will shed light upon some of these mysteries. I will employ diverse ideas---from thermodynamics and optimal transportation to partial differential equations, control theory and Bayesian inference---and paint a picture of the training process of deep networks. Along the way, I will develop state-of-the-art algorithms for non-convex optimization.

The goal of machine perception is not just to classify objects in images but instead, enable intelligent agents that can seamlessly interact with our physical world. I will conclude with a vision of how advances in machine learning and robotics may come together to help build such an Embodied Intelligence.

Pratik Chaudhari is an Applied Scientist at Amazon Web Services and a post-doctoral scholar at Caltech. He will be joining the faculty in the Electrical Engineering department at the University of Pennsylvania in Fall 2019. He holds a PhD in Computer Science from UCLA and an MS in Aero-Astro from MIT. His research interests include deep learning, robotics and computer vision. Pratik has worked on perception and control algorithms for safe autonomous urban driving as a part of nuTonomy Inc.