[Computer-go] MCTS for Games Special Session

[Lucas, Simon M]

Hi all,

The following special session held as part of
IEEE CIG 2012 in Granada (September) may be of interest.

Best wishes,

  Simon Lucas



Monte Carlo Tree Search for Games
http://geneura.ugr.es/cig2012/sessions.html

Monte Carlo Tree Search (MCTS) is a novel tree search algorithm that has led to recent breakthroughs in computer Go and other games. MCTS, unlike classical tree search algorithms, is any-time and does not necessarily require a state value function. These attributes make MCTS a promising candidate for games in which traditional tree search algorithms tend to fail. MCTS has been applied successfully to a wide variety of games, including puzzles, board games, card games and real-time video games, and has produced world champion AI players for a number of difficult games. More recently, MCTS has also been proposed as a method for automated level and game design.
The aim of this special session is to highlight these advances and to bring together high quality publications concerning the study, analysis and application of MCTS with respect to games.

Topics of interest include, but are not limited to:
MCTS for intelligent move planning
MCTS for opponent modelling
Learning in MCTS
Hybrid MCTS algorithms
MCTS for real-time (video) games
MCTS for continuous domains
MCTS for open-ended (infinite) domains
MCTS for nondeterministic / hidden information games
MCTS for social (multiplayer / euro / bidding / gambling) games
MCTS for solitaire puzzle design and solving
MCTS for game, content and level design
Enhancements and variations of MCTS


organizers: Philipp Rohlfshagen, Edward Powley and Cameron Browne


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://computer-go.org/pipermail/computer-go/attachments/20120306/dba043fb/attachment.html>