BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning\, Planning and Representing Knowledge from Primitive Expe
 rience - David Silver - UCL
DTSTART:20100413T130000Z
DTEND:20100413T140000Z
UID:TALK24129@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:*Abstract:* Sequential decision-making and planning problems a
 re ubiquitous in engineering\, economics\, finance\, artificial intelligen
 ce and robotics. Many of the hardest and most important problems involve l
 ong-term consequences of immediate choices\, and therefore vast search-spa
 ces. This talk proposes a new paradigm for general-purpose planning and de
 cision-making that is appropriate for these large and challenging problems
 . The inspiration for this approach comes from the game of Go. This task i
 s widely viewed as a grand challenge for artificial intelligence\, which h
 as thwarted traditional approaches to knowledge representation\, learning 
 and planning. Recently\, a radically different approach has been developed
 \, resulting in the first program\, MoGo\, to perform at human master leve
 l. This revolution in computer Go is based on three key principles: a) kno
 wledge is represented by predictions about future experience\, b) predicti
 ons are learnt directly from experience\, c) planning is achieved by simul
 ating experience. This talk proposes a more general application of this ex
 perience-based strategy. Just like Go\, many real-world planning and decis
 ion-making problems have enormous search-spaces that are intractable to tr
 aditional search algorithms. Furthermore\, also just like Go\, in the majo
 rity of these problems\, expert knowledge is unavailable or unreliable. 
LOCATION:Small public lecture room\, Microsoft Research Ltd\, 7 J J Thomso
 n Avenue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR