BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250810T231945EDT-0329zCsZo9@132.216.98.100
DTSTAMP:20250811T031945Z
DESCRIPTION:Title: Variance reduction by occluding a Markov chain.\n\nAbstr
 act:\n\nStochastic algorithms which simulate random variables/processes on
  a computer to estimate intractable quantities are ubiquitous in Statistic
 s and elsewhere. One such method is Markov chain Monte Carlo which\, under
  mild conditions\, offer asymptotical (in time) guarantees. In this talk\,
  we define infinitely many stopping times at which an ergodic Markov chain
  is occluded by a (conditionally) independent process. The resulting proce
 ss\, called the occluded process\, is not Markov\, but provided that the s
 topping times/independent process are cleverly defined\, we show that it i
 s ergodic. One particularly powerful way to define the stopping times/inde
 pendent process leverages the recent advances in ML regarding approximatio
 ns of probability distributions (divergence minimization\, normalizing flo
 ws\, etc.). We discuss the variance reduction effect of the occluded proce
 ss through some illustrations and (weak) theoretical results in some limit
 ing regime.\n\nSpeaker\n\nFlorian Maire is an Associate Professor at Unive
 rsité de Montréal\, which he joined in 2018 after a PhD from Paris 6 and T
 elecom SudParis and a Postdoc at University College Dublin. His research i
 nterests are mostly in computational statistics and stochastic optimizatio
 n.\n
DTSTART:20240920T193000Z
DTEND:20240920T203000Z
LOCATION:Room 1104\, Burnside Hall\, CA\, QC\, Montreal\, H3A 0B9\, 805 rue
  Sherbrooke Ouest
SUMMARY:Florian Maire(Université de Montréal)
URL:/channels/channels/event/florian-maireuniversite-d
 e-montreal-359670
END:VEVENT
END:VCALENDAR