BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Information Constrained Optimal Transport
DTSTART:20230706T110000
DTEND:20230706T120000
DTSTAMP:20260406T194655Z
UID:be078dca1bb8c9f8cdac64352b07eed66c8892048fc30394e11341e5
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Ayfer Özgür\n\nOptimal transport is the general proble
 m of moving one distribution of mass to another as efficiently as possible
 . This problem has found a wealth of applications within mathematics as we
 ll as many other fields. In this talk\, we will introduce an information c
 onstrained version of optimal transport and show that 1) it leads to the s
 olution of an open problem in information theory formulated by Cover in 19
 87 which he calls the capacity of the relay channel\; 2) it removes the cu
 rse of dimensionality in Wasserstein GANs and can be used to optimally tra
 in generative models from privatized data.\n\nProf. Ozgur is an associate 
 professor in the Electrical Engineering Department of Stanford University 
 and a member of ISL. Before joining Stanford\, she was a postdoctoral rese
 archer with the Algorithmic Research on Networked Information Group at EPF
 L\, Switzerland. She received my PhD degree in October\, 2009 from the Inf
 ormation Processing Group at the same university.\n\n\n \n\n 
LOCATION:INR 113 https://plan.epfl.ch/?room==INR%20113
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR