BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Colloquium: Improving prophet inequalities: going beyond worst-
 case and the impact of weaker benchmark
DTSTART:20260629T101500
DTEND:20260629T111500
DTSTAMP:20261002T062227Z
UID:db00273006be8f5233090709cb8f6a2d82c88df53c42561b8c8adede
CATEGORIES:Conferences - Seminars
DESCRIPTION:By: Vianney Perchet - CREST at ENSAE\nVideo of the talk\n\nAbs
 tract\nProphet inequalities is a standard stopping time question where a d
 ecision maker observes sequentially random variables and decides when to s
 top\, in the hope of maximizing the value of the selected variable. There 
 is an extensive literature on the subject\, proving some "optimality" of d
 ifferent procedures in the worst-case scenario. We claim\, and prove\, tha
 t those results can be strikingly improved when considering non-pathologi
 cal distributions. This is illustrated in two directions: by introducing 
 a parameter that characterizes the complexity of an instance or by looking
  at some slightly weaker benchmarks that strongly impact the performances
  of the oracle that knows in advance all the realized values.\n\nBio\nVian
 ney Perchet is a professor at the Centre de recherche en économie et stat
 istique (CREST) at the ENSAE since october 2019. Mainly focusing on the in
 terplay between machine learning and game theory\, his research themes are
  at the intersection of mathematics\, computer science\, and economics. Th
 e spectrum of his interest ranges from pure theory  (say\, optimal rates 
 of convergence of algorithms) to pure applications (modeling user behavior
 \, optimisation of recommender systems\, etc.) He is also part-time distin
 guished researcher in the Criteo AI Lab\, in Paris\, working on efficient 
 exploration in recommender systems.\n\nMore information
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420 https://epfl.zoom.us/
 j/66307507932
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR