BEGIN:VCALENDAR
VERSION:2.0
PRODID:ILLC Website
X-WR-TIMEZONE:Europe/Amsterdam
BEGIN:VTIMEZONE
TZID:Europe/Amsterdam
X-LIC-LOCATION:Europe/Amsterdam
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700329T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701025T030000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:/NewsandEvents/Archives/2024/newsitem/14815/16
 -February-2024-Formalisation-Optimisation-Algorith
 ms-Mechanisms-FOAM-Matthijs-Spaan
DTSTAMP:20240212T145900
SUMMARY:Formalisation, Optimisation, Algorithms, M
 echanisms (FOAM), Matthijs Spaan
ATTENDEE;ROLE=Speaker:Matthijs Spaan
DTSTART;TZID=Europe/Amsterdam:20240216T150000
DTEND;TZID=Europe/Amsterdam:20240216T162500
LOCATION:Room L3.33, ILLC Lab42, Science Park 900,
  Amsterdam
DESCRIPTION:In this talk I discuss how estimating 
 and propagating epistemic uncertainty benefits gen
 eralization and deep exploration in reinforcement 
 learning (RL) by focusing on two recent contributi
 ons. First, I consider model-free distributional R
 L, which aims to learn the distribution of returns
  rather than their expected value. Second, I discu
 ss how propagating epistemic uncertainty estimates
  can be leveraged in a model-based RL setting, by 
 embedding them in Monte-Carlo Tree Search (MCTS).
X-ALT-DESC;FMTTYPE=text/html:\n  <p>In this talk I
  discuss how estimating and propagating epistemic 
 uncertainty benefits generalization and deep explo
 ration in reinforcement learning (RL) by focusing 
 on two recent contributions. First, I consider mod
 el-free distributional RL, which aims to learn the
  distribution of returns rather than their expecte
 d value. Second, I discuss how propagating epistem
 ic uncertainty estimates can be leveraged in a mod
 el-based RL setting, by embedding them in Monte-Ca
 rlo Tree Search (MCTS).</p>\n
URL:https://events.illc.uva.nl/FOAM/posts/talk11/
CONTACT:Gregor Behnke at g.behnke at uva.nl
CONTACT:Ronald de Haan at r.dehaan at uva.nl
END:VEVENT
END:VCALENDAR