BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:FLAIR external seminar: Universal coding under Gaussian noise\, an
 d the Wills functional
DTSTART:20230317T131500
DTEND:20230317T141500
DTSTAMP:20260407T103358Z
UID:2ac0d14899bd7411b79a9476deeac3cd9fcada1771f00d0781d9809f
CATEGORIES:Conferences - Seminars
DESCRIPTION:Jaouad Mourtada\nSequential probability assignment is a classi
 cal prediction problem\, wherein one aims to assign a large probability to
  a sequence of observations revealed one at a time. This problem is closel
 y related to that of lossless data compression (also called universal codi
 ng) in information theory. We study this problem in the case where the bas
 e model is a subset of the standard Gaussian model\, with mean constrained
  to a general convex domain. We show that the hardness of the problem can 
 be expressed in terms of certain quantities from convex geometry\, namely 
 the intrinsic volumes of the constraint set. We then provide an alternati
 ve characterization of the optimal error in terms of metric complexity mea
 sures\, which also holds in the general nonconvex case. We finally relate
  and contrast our findings with classical asymptotic results in informatio
 n theory.\n 
LOCATION:GA 3 21 https://plan.epfl.ch/?room==GA%203%2021
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR