BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:State Space Abstraction for Reinforcement Learning - Rowan McAllis
 ter\; Thang Bui
DTSTART:20141106T150000Z
DTEND:20141106T163000Z
UID:TALK55674@talks.cam.ac.uk
CONTACT:39777
DESCRIPTION:Reinforcement learning (RL) is a method of solving sequential 
 decision tasks in Markov decision process frameworks with unknown paramete
 rs. Unfortunately the computational complexity of RL hinders application t
 o many real world problems. Function approximation techniques are a common
  method to compute approximate solutions fast. An alternative technique is
  state space abstraction\, discarding irrelevant state information. Abstra
 cted state spaces speed up learning exponentially w.r.t. state dimensional
 ity. An extra benefit is discovery of powerful generalisations in the orig
 inal state space. This talk will provide a review of state space abstracti
 on. We introduce different types of abstractions and their consequences to
  the solution accuracy. We will also discuss predictive state representati
 ons: a compact way to model dynamical systems using predictions of observa
 ble quantities.
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR