We are delighted to announce that the Simons Institute will play a key role in a new project funded by the National Science Foundation (NSF) and the Simons Foundation devoted to investigating the theoretical foundations of deep learning.
NSF and the Simons Foundation have awarded $10 million to this five-year project, which will be led by Simons Institute Associate Director Peter Bartlett. Most of the project's research and education activities, including collaborative research meetings and workshops, will be hosted by the Simons Institute.
Deep learning is part of a broader family of machine learning methods based on artificial neural networks that digest large amounts of raw data and train artificial intelligence systems with limited human supervision. Although deep learning is a widely used AI approach for teaching computers to learn from data, its theoretical foundations are poorly understood — a challenge that the project will address. Understanding the mechanisms that underpin the practical success of deep learning will allow researchers to address its limitations, including its sensitivity to data manipulation.
"Our excitement over receiving this award is that we will be digging into the theoretical foundations of deep learning," Bartlett remarked. "The recent success in machine learning has been driven by a spirit of craftsmanship by people who find ways to make this technology successful. But much of this work contradicts a lot of our classical understanding of statistical methodology, and there are many things we don't understand about how and why these systems work."
The other leaders of this Berkeley-led project are Bin Yu (UC Berkeley), Andrea Montanari (Stanford), Emmanuel Abbe (EPFL), Misha Belkin (UC San Diego), Amit Daniely (Hebrew University), Sasha Rakhlin, Elchanan Mossel and Nike Sun (MIT), Nati Srebro (TTIC), and Roman Vershynin (UC Irvine).
Read the full press release from UC Berkeley's new Division of Computing, Data Science, and Society (CDSS).