Expected-Outcome: A General Model of Static Evaluation
IEEE Transactions on Pattern Analysis and Machine Intelligence
Machines that learn to play games
Strategies for the Automatic Construction of Opening Books
CG '00 Revised Papers from the Second International Conference on Computers and Games
CG '08 Proceedings of the 6th international conference on Computers and Games
IJCAI'09 Proceedings of the 21st international jont conference on Artifical intelligence
Efficient selectivity and backup operators in Monte-Carlo tree search
CG'06 Proceedings of the 5th international conference on Computers and games
Bandit based monte-carlo planning
ECML'06 Proceedings of the 17th European conference on Machine Learning
Evaluation function based monte-carlo LOA
ACG'09 Proceedings of the 12th international conference on Advances in Computer Games
Creating an upper-confidence-tree program for havannah
ACG'09 Proceedings of the 12th international conference on Advances in Computer Games
Innovative opening-book handling
ACG'05 Proceedings of the 11th international conference on Advances in Computer Games
Generating an opening book for amazons
CG'04 Proceedings of the 4th international conference on Computers and Games
| Hi-index | 0.00 |
Automatically creating opening books is a natural step towards the building of strong game-playing programs, especially when there is little available knowledge about the game. However, while recent popular Monte-Carlo Tree-Search programs showed strong results for various games, we show here that programs based on such methods cannot efficiently use opening books created using algorithms based on minimax. To overcome this issue, we propose to use an MCTS-based technique, Meta-MCTS, to create such opening books. This method, while requiring some tuning to arrive at the best opening book possible, shows promising results to create an opening book for the game of the Amazons, even if this is at the cost of removing its Monte-Carlo part.