Static analysis of life and death in the game of Go
Information Sciences—Informatics and Computer Science: An International Journal
Computer Go: an AI oriented survey
Artificial Intelligence
Artificial Intelligence - Chips challenging champions: games, computers and Artificial Intelligence
Finite-time Analysis of the Multiarmed Bandit Problem
Machine Learning
Feature extraction and representation for pattern recognition and the game of go
Feature extraction and representation for pattern recognition and the game of go
Heuristic analysis of large trees as generated in the game of 'go'
Heuristic analysis of large trees as generated in the game of 'go'
Combining online and offline knowledge in UCT
Proceedings of the 24th international conference on Machine learning
A Fast Indexing Method for Monte-Carlo Go
CG '08 Proceedings of the 6th international conference on Computers and Games
Algorithms and application in decision-making for the finest splitting of a set of formulae
Knowledge-Based Systems
Associating domain-dependent knowledge and Monte Carlo approaches within a Go program
Information Sciences: an International Journal
Explaining how to play real-time strategy games
Knowledge-Based Systems
A rough set approach to feature selection based on power set tree
Knowledge-Based Systems
Dynamic Randomization Enhances Monte-Carlo Go
TAAI '10 Proceedings of the 2010 International Conference on Technologies and Applications of Artificial Intelligence
The 10th International Conference on Autonomous Agents and Multiagent Systems - Volume 1
Bandit based monte-carlo planning
ECML'06 Proceedings of the 17th European conference on Machine Learning
Adding expert knowledge and exploration in monte-carlo tree search
ACG'09 Proceedings of the 12th international conference on Advances in Computer Games
Bitboard knowledge base system and elegant search architectures for Connect6
Knowledge-Based Systems
| Hi-index | 0.00 |
This paper is an extension of the article [13] presented at IWCG of TAAI 2010. It proposes two dynamic randomization techniques for Monte-Carlo Tree Search (MCTS) in Go. First, during the in-tree phase of a simulation game, the parameters are randomized in selected ranges before each simulation move. Second, during the play-out phase, the priority orders of the simulation move-generators are hierarchically randomized before each play-out move. Essential domain knowledge used in MCTS for Go is discussed. Both dynamic randomization techniques increase diversity while keeping the sanity of the simulation games. Experimental testing has been completely re-conducted more extensively with the latest version of GoIntellect (GI) on all three Go categories of 19x19, 13x13, and 9x9 boards. The results show that dynamic randomization increases the playing strength of GI significantly with 128K simulations per move, the improvement is about seven percentage points in the winning rate against GnuGo on 19x19 Go over the version of GI without dynamic randomization, about three percentage points on 13x13 Go, and four percentage points on 9x9 Go.