The National Science Foundation (NSF) and Simons Foundation Division of Mathematics and Physical Sciences recently announced a 5-year, $10M award for the Collaboration on the Theoretical Foundations of Deep Learning. Berkeley EECS professor and BIDS Faculty Affiliate Bin Yu will be a co-PI on the project, along with researchers from Stanford University, MIT, UC Irvine, UC San Diego, the Toyota Technological Institute at Chicago, EPFL in Lausanne, Switzerland, and the Hebrew University in Jerusalem. The interdisciplinary team will be led by the Associate Director of Berkeley’s Simons Institute for the Theory of Computing, Peter Bartlett.
Deep learning is part of a broader family of machine learning methods that uses artificial neural networks to analyze large amounts of raw data and train artificial intelligence systems, usually with limited or no direct human supervision. The Berkeley-led team will investigate the theoretical foundations and underlying mechanisms of deep learning algorithms, which will allow researchers to understand and address its limitations and sensitivities.
Especially with applications in human health and medicine, it is crucial to understand how deep learning works, and to confirm that the algorithms being applied to specialized datasets lead to realistic and accurate results. “This is really cutting-edge research at the frontier of data science theory and practice,” according to Yu. “It’s leading us intellectually where we want to go with deep learning. Theory can help improve our practice, and that practice then drives theory.”
Read more...
UC Berkeley to lead $10M NSF/Simons Foundation program to investigate theoretical underpinnings of deep learning
August 25, 2020 | Jon Bashor | CDSS News
NSF and the Simons Foundation partner to uncover foundations of deep learning: Mathematicians, statisticians, engineers and theoretical computer scientists come together to work on deep learning and AI
August 25, 2020 | NSF Press Announcement
