"Artificial Neural Net Othello"
Robert Hoff 2004
Abstract
This Independent Study explores the history and techniques of automated game playing, focusing on a learning program developed for checkers. Software to play Othello is created using an artificial neural network evaluation function. An attempt is made to train the network by back-propagation on data from a log of existing Othello games. The goal is a network which is successful against difficult opponents without the need to search more than one move ahead. The chosen network structure proved too constrictive for the goal behavior to emerge.