- This event has passed.
Austin STROMME (MIT) – “Statistical optimal transport: barycenters and entropic regularization”
Statistical Seminar: Every Monday at 2:00 pm.
Time: 2:00 pm – 3:15 pm
Date: 6th of March 2023
Place: Room 3001 + ZOOM
Austin STROMME – “Statistical optimal transport: barycenters and entropic regularization”
https://zoom.us/j/98407728890?pwd=aEgzaVJvd0xCZEZnTUg5emtIRGJYZz0
Abstract:
Optimal transport (OT) is a popular paradigm for comparing and interpolating datasets using a minimum energy criterion, and gives rise to a Riemannian-like geometry on probability distributions. In this talk, we’ll study statistical and connected algorithmic questions relating to some of the most promising facets of OT: barycenters, OT on Gaussians, entropic regularization, and transfer learning. Throughout, we’ll emphasize dimension-free rates of estimation and practical algorithms, as well as the geometry and optimization-centric viewpoint that yields such results.
Organizers:
Cristina BUTUCEA (CREST), Alexandre TSYBAKOV (CREST), Karim LOUNICI (CMAP) , Jaouad MOURTADA (CREST)
Sponsors:
CREST-CMAP