BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250712T092211EDT-4328Wku2hF@132.216.98.100
DTSTAMP:20250712T132211Z
DESCRIPTION:Safe Policy Improvement with Baseline Bootstrapping\n\n A commo
 n goal in Reinforcement Learning is to derive a good strategy given a limi
 ted batch of data. In this paper\, we propose a new strategy to compute a 
 safe policy\, guaranteed to perform at least as well as a given baseline s
 trategy. We advocate that the assumptions made in previous work are too st
 rong for real world applications and propose new algorithms allowing those
  assumptions to be satisfied only in a subset of the state-action pairs. W
 hile significantly relaxing the assumptions\, our algorithms achieve the s
 ame accuracy guarantees than the previous work\, and are also much more co
 mputationally efficient. We also show that the algorithms can be adapted t
 o model-free Reinforcement Learning. \n
DTSTART:20171201T153000Z
DTEND:20171201T163000Z
LOCATION:Room AA6214\, CA\, Pav. André-Aisenstadt
SUMMARY:Romain Laroche\, researcher from MS Maluuba
URL:/mathstat/channels/event/romain-laroche-researcher
 -ms-maluuba-283119
END:VEVENT
END:VCALENDAR