Learning on Graphs in the Game of Go
We consider the game of Go from the point of view of machine learning and as a well-defined domain for learning on graph representations. We discuss the representation of both board positions and candidate moves and introduce the common fate graph (CFG) as an adequate representation of board positions for learning. Single candidate moves are represented as feature vectors with features given by subgraphs relative to the given move in the CFG. Using this representation we train a support vector machine (SVM) and a kernel perceptron to discriminate good moves from bad moves on a collection of life-and-death problems and on 9 × 9 game records. We thus obtain kernel machines that solve Go problems and play 9 × 9 Go.
Unable to display preview. Download preview PDF.
- 2.J. Burmeister and J. Wiles. The challenge of go as a domain for ai research: A comparison between go and chess. In Proceedings of the 3rd Australian and New Zealand Conference on Intelligent Information Systems, 1994.Google Scholar
- 4.D. Fotland. Knowledge representation in The Many Faces of Go, 1993.Google Scholar
- 5.P. Geibel and F. Wysotzki. Learning relational concepts with decision trees. In Machine Learning: Proceedings of the Thirteenth International Conference, pages 1141–1144. Morgan Kaufmann Publishers, 1998.Google Scholar
- 6.N. Sasaki and Y. Sawada. Neural networks for tsume-go problems. In Proceedings of the Fifth International Conference on Neural Information Processing, pages 1141–1144, 1998.Google Scholar
- 7.N. N. Schraudolph, P. Dayan, and T. J. Sejnowski. Temporal difference learning of position evaluation in the game of go. In J. D. Cowan, G. Tesauro, and J. Alspector, editors, Advances in Neural Information Processing Systems, volume 6, pages 817–824. Morgan Kaufmann Publishers, Inc., 1994.Google Scholar
- 9.T. Wolf. The program GoTools and its computer-generated Tsume Go database. In Proceedings of the 1st Game Programming Workshop, 1994.Google Scholar