Speaker: Chenyang Zhong, Columbia University, Department of Statistics
Title: "Entropic Selection in Optimal Transport"
Abstract:
Optimal transport is ubiquitous across mathematics and its applications, impacting fields from analysis and geometry to image processing and machine learning. Thanks to efficient implementation through Sinkhorn’s algorithm, entropic regularization has become a key computational method for solving optimal transport in high dimensions. While the original optimal transport problem can admit many solutions—as is the case for the Euclidean distance cost on R^d (Monge’s problem)—entropically regularized optimal transport (EOT) always has a unique optimizer. A longstanding open question has been whether, as the regularization parameter vanishes, the EOT solution converges to a distinguished optimal transport coupling, and if so, what that coupling looks like. In this talk, I will present a surprising resolution of this problem for Monge’s problem in dimension d>1. This is joint work with Marcel Nutz.
Bio:
Chenyang Zhong is an Assistant Professor in the Department of Statistics at Columbia University. He received his Ph.D. in Statistics from Stanford University, advised by Persi Diaconis. His research lies at the intersection of applied probability and statistics, with a focus on Markov chain Monte Carlo, variational inference, optimal transport, and random matrix theory.
In person attendance at this seminar is only open to Columbia University affiliates. External guests are welcome to attend remotely. Please contact [email protected] if you need the Zoom link for this seminar.
