BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Coherent Inference on Optimal Play in Games - Philipp Hennig (Univ
 ersity of Cambridge)
DTSTART:20100510T100000Z
DTEND:20100510T110000Z
UID:TALK24617@talks.cam.ac.uk
CONTACT:Emli-Mari Nel
DESCRIPTION:The search for an optimal path through a game tree is one of t
 he oldest problems in computer science. Over the past years\, Monte Carlo 
 tree search has emerged as a surprisingly effective approach to this probl
 em. I will present a probabilistic generative model for game trees that ex
 plains why MC tree search works at all. I will then move on to derive an a
 pproximate inference algorithm for this model\, which can infer beliefs ov
 er the value of _any_ node in the tree under _optimal_ play\, using random
  roll-out data from other parts in the tree. Somewhat surprisingly\, this 
 inference algorithm is of linear complexity\, even though the exact search
  problem has exponential cost.\n\nThe work presented in this talk has just
  been published as P. Hennig\, D. Stern\, T. Graepel: "Coherent Inference 
 on Optimal Play in Game Trees"\, J Machine Learning Research\, W&CP 9 (201
 0)\, 326-333. \n\nSee\nhttp://jmlr.csail.mit.edu/proceedings/papers/v9/hen
 nig10a/hennig10a.pdf
LOCATION:TCM Seminar Room\, Cavendish Laboratory\, Department of Physics
END:VEVENT
END:VCALENDAR