From Classical Statistics to Modern ML: the Lessons of Deep Learning
Mikhail Belkin
Ohio State University
October 16, 2019
School of Mathematics
Spotlight talks
Various
October 16, 2019
Towards a theoretical foundation of neural networks
Jason Lee
Princeton University; Member, School of Mathematics
October 16, 2019
Panel Session
Various
October 16, 2019
Energy-based Approaches to Representation Learning
Yann LeCun
NYU and Facebook AI
October 16, 2019
Spotlight Talks
Yuanzhi Li, Soham De, Mahyar Fazlyab, Maithra Raghu, Valentin Thomas
October 15, 2019
Asymptotic spectra and Applications II
Jeroen Zuiddam
Member, School of Mathematics
October 15, 2019
In this second lecture in my series on asymptotic spectra we will focus on one application: the matrix multiplication problem. We will use the asymptotic spectrum of tensors to prove that a very general method (that includes the methods used to obtain the currently best algorithms) cannot give faster matrix multiplication algorithms. Keywords are Shannon entropy, representation theory and moment polytopes, but prior knowledge of these is not assumed.
Emergent linguistic structure in deep contextual neural word representations
Chris Manning
Stanford University
October 15, 2019
What 2-layer neural nets can we optimize?
Rong Ge
Duke University; Member, School of Mathematics
October 15, 2019
Tightening information-theoretic generalization bounds with data-dependent estimates with an application to SGLD
Daniel Roy
University of Toronto
October 15, 2019