Learning Middle-Game Patterns in Chess: A Case Study
Abstract
Despite the undisputed strength of today’s chess-playing programs, the fact that they have to evaluate millions, or even billions, of different positions per move is unsatisfactory. The amount of “computation” carried out by human players is smaller by orders of magnitudes because they employ specific patterns that help them narrow the search tree. Similar approachs hould in principle be feasible also in computer programs. To draw attenion to this issue, we report our experiments with a program that learns to classify chessboard positions that permit the well-known bishop sacrifice at h7. We discuss some problems pertaining to the collection of training examples, their representation, and pre-classification. Classification accuracies achieved with a decision-tree based classifier are encouraging.
Keywords
computer chess pattern recognition concept learningPreview
Unable to display preview. Download preview PDF.
References
- 1.deGroot, A. D. (1965). Thought and Choice in Chess, Mouton, The HagueGoogle Scholar
- 2.Flinter, S. and Keane, M. T. (1995). Using Chunking for the Automatic Generation of Cases in Chess. Proceedings of the 1st International Conference on Case Based Reasoning, ICCBR-95, Springer VerlagGoogle Scholar
- 3.Fürnkranz, J. (1996). Machine Learning in Computer Chess: The Next Generation. International Computer Chess Association Journal, 19, 147–160Google Scholar
- 4.Gobet, J. and Jansen, P. (1994). Towards a Chess Program Based on a Model of Human Memory. In H. J. van den Herik et al. (eds.) Advances in Computer Chess, 7, pp. 35–60, University of LimburgGoogle Scholar
- 5.Kubat, M. and Matwin, S. (1997). Addressing the Curse of Imbalanced Training Sets: One-Sided Selection. —em Proceedings of the Fourteenth International Conference ICML’97. July 8–12, 1997, Nashville, Tennessee, pp. 179–186.Google Scholar
- 6.Mitchell, T. M. (1996). Machine Learning, McGraw HillGoogle Scholar
- 7.Quinlan, J. R. (1996). Bagging, Boosting, and C4.5. Proceedings of the Eight Annual Conference on Innovative Applications of Artificial Intelligence, AAAI’96. August 4–8, 1996, Portland, Oregon, pp. 725–730.Google Scholar
- 8.Shannon, C. E. (1950). Programming a Computer for Playing Chess. Philosophical Magazine, 41(4), 256–275MATHMathSciNetGoogle Scholar
- 9.Shapiro, A. D. (1987). Structured Induction in Expert Systems. Turing Institute Press, Addison-WesleyGoogle Scholar
- 10.Weill, J.-C. (1994). How Hard is the Correct Coding of an Easy Endgame. In H. J. van den Herik et al. (eds.) Advances in Computer Chess, 7, pp. 163–176, University of LimburgGoogle Scholar