BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Cooperative Inverse Reinforcement Learning - Robert Pinsler (Unive
 rsity of Cambridge)
DTSTART:20171025T160000Z
DTEND:20171025T173000Z
UID:TALK100036@talks.cam.ac.uk
CONTACT:Adrià Garriga Alonso
DESCRIPTION:The value alignment problem consists in ensuring the values of
  an AI system align with the values of its operator. A potential solution 
 to this problem is\nformalised as the Inverse Reinforcement Learning (IRL)
  setting. In IRL\, the goal is to infer the reward function of an agent (a
  human)\, just from observing its\nbehaviour in the environment. In Cooper
 ative IRL\, the agents are allowed to interact. From this\, more effective
  teaching strategies than passive observation\nemerge. We will talk about 
 formalising this problem\, and an algorithm to approximate good teaching s
 trategies.\n\nTalk slides: https://valuealignment.ml/talks/2017-10-25-coop
 erative-inverse-reinforcement-learning.pdf
LOCATION: Cambridge University Engineering Department\, CBL Seminar room B
 E4-38.  For directions see http://learning.eng.cam.ac.uk/Public/Directions
END:VEVENT
END:VCALENDAR