BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Proximal optimal transport divergences for generative modeling and
  sampling
DTSTART:20261111T150000
DTEND:20261111T160000
DTSTAMP:20261109T145510Z
UID:deef33c8aa84f3336fe1ac64e509c12bccae308c4b43ed128abe4c25
CATEGORIES:Conferences - Seminars
DESCRIPTION:Professor Luc Rey-Bellet\, UMass Amherst\nThe motivation for t
 his work comes from generative modeling and sampling problems where the ta
 rget distributions are either supported on low-dimensional structure and/o
 r known only through empirical samples.  Through proximal regularization 
 we construct new information theoretic divergences which combine in a flex
 ible manner the desirable properties  of relative entropy  and optimal t
 ransport (Wasserstein distances).  Using these divergences\, associated W
 asserstein gradient flows\, and a mixture of implicit and explicit schemes
  we build efficient sampling and generative modeling algorithms.
LOCATION:GA 3 21 https://plan.epfl.ch/?room==GA%203%2021
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR