Gobet F. & Jansen, P. (1994).

Towards a chess program based on a model of human memory. In H. J. van den Herik, I. S. Herschberg, & J. W.Uiterwijk (Eds.), Advances in Computer Chess 7. Maastricht: University of Limburg Press.

 

Abstract

Research in cognitive psychology has shown that pattern recognition represents a key element of chess skill in humans. Because of the availability of powerful computer search techniques, this approach has been largely neglected by the computer-chess community. In this paper, we present a chess program, CHUMP, that learns patterns of pieces, associates moves and move sequences with these patterns, and subsequently uses this knowledge to suggest moves. The program's proposed moves are compared with actual moves in several different learning and testing situations (e.g., games by Mikhail Tal and positions from the KQKR domain). The experiments show that CHUMP improves performance with learning in all cases. We discuss the relevance of our method, as well as the modifications and extensions necessary to obtain a competitive chess program.