A Methodology for Learning Players' Styles from Game Records - Computer Science > Artificial IntelligenceReport as inadecuate




A Methodology for Learning Players' Styles from Game Records - Computer Science > Artificial Intelligence - Download this document for free, or read online. Document in PDF available to download.

Abstract: We describe a preliminary investigation into learning a Chess player-s stylefrom game records. The method is based on attempting to learn features of aplayer-s individual evaluation function using the method of temporaldifferences, with the aid of a conventional Chess engine architecture. Someencouraging results were obtained in learning the styles of two recent Chessworld champions, and we report on our attempt to use the learnt styles todiscriminate between the players from game records by trying to detect who wasplaying white and who was playing black. We also discuss some limitations ofour approach and propose possible directions for future research. The method wehave presented may also be applicable to other strategic games, and may even begeneralisable to other domains where sequences of agents- actions are recorded.



Author: Mark Levene, Trevor Fenner

Source: https://arxiv.org/



DOWNLOAD PDF




Related documents