| inria-00117266, version 3 |
|
|
| Voir la fiche détaillée | BibTeX EndNote TEI RefWorks |
|
|
|||||||
| Algorithm UCB1 for multi-armed bandit problem has already been extended to Algorithm UCT (Upper bound Confidence for Tree) which works for minimax tree search. We have developed a Monte-Carlo Go program, MoGo, which is the first computer Go program using UCT. We explain our modification of UCT for Go application and also the intelligent random simulation with patterns which has improved significantly the performance of MoGo. UCT combined with pruning techniques for large Go board is discussed, as well as parallelization of UCT. MoGo is now a top level Go program on $9\times9$ and $13\times13$ Go boards. |
|
|
|
|
|
|
|
|
| 1 : | TAO (INRIA Futurs) |
| INRIA – CNRS : UMR8623 – Université Paris Sud - Paris XI | |
| 2 : | Centre de Mathématiques Appliquées (CMAP) |
| CNRS : UMR7641 – Université de Versailles-Saint Quentin en Yvelines – Polytechnique - X | |
| 3 : | SEQUEL (INRIA Futurs) |
| INRIA – CNRS : UMR8022 – CNRS : UMR8146 – Université des Sciences et Technologies de Lille - Lille I – Université Charles de Gaulle - Lille III – Ecole Centrale de Lille |
|
|
|
|
|
|
|
|
| Domaine | : | Informatique/Intelligence artificielle |
| RR-6062 |
| Versions disponibles : | v1 (30-11-2006) | v2 (12-12-2006) | v3 (21-12-2006) |
| inria-00117266, version 3 | |
| http://hal.inria.fr/inria-00117266/fr/ | |
| oai:hal.inria.fr:inria-00117266_v3 | |
| Contributeur : Sylvain Gelly | |
| Soumis le : Mercredi 20 Décembre 2006, 18:52:43 | |
| Dernière modification le : Jeudi 21 Décembre 2006, 09:49:17 | |