BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Instant-Optimal Algorithms for Pure Exploration in Reinforcement L
 earning
DTSTART:20251202T110000
DTEND:20251202T120000
DTSTAMP:20260407T111406Z
UID:a8672d316bf9b0ebb6b58b8fe2823e27fec7b6c3dc3bfeaadbf3cedc
CATEGORIES:Conferences - Seminars
DESCRIPTION:Cyrille Kone\, PhD\, University of Lille\nAbstract\nInstant-Op
 timal Algorithms for Pure Exploration in Reinforcement Learning\nIn online
  reinforcement learning\, pure exploration aims to identify an optimal pol
 icy after a learning phase with minimal sample complexity\, in contrast to
  regret minimization which focuses on performance during learning. We stud
 y instance-dependent lower bounds for this problem\, which take the form o
 f a two-player zero-sum game between an explorer choosing a behavior polic
 y and nature selecting an alternative MDP. We propose a computationally ef
 ficient algorithm based on posterior sampling that matches this lower boun
 d in the small-error regime\, bypassing the hardness of computing best res
 ponses. We further discuss extensions to multi-agent reinforcement learnin
 g\, where the goal is to identify strategic equilibria such as Nash equili
 bria in unknown environments.\n\nBiography\nCyrille Kone is a PhD candidat
 e in Computer Science at the University of Lille and Inria within the Scoo
 l team\, supervised by Prof. Emilie Kaufmann and Prof. Laura Richert. His 
 research focuses on the theoretical foundations of sequential decision-mak
 ing\, with emphasis on pure exploration in bandits and reinforcement learn
 ing\, instance-optimal algorithm design\, and multi-objective optimization
 . His work has been published at top machine learning venues including Neu
 rIPS\, ICML\, and AISTATS. He will defend his PhD in December 2025. 
LOCATION:ME C2 405 https://plan.epfl.ch/?room==ME%20C2%20405
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR