AI techniques for the game of Go
Erik van der Werf · 2005
In the last decade Go has been an important part of my life.As a student in Delft I became fascinated by the question why, unlike Chess, computers played this game so poorly.This fascination stimulated me to pursue computer Go as a hobby and I was fortunate to share my interests with some fellow students with whom I also founded a small Go club.In the final years of my study applied physics I joined the pattern recognition group where I performed research on non-linear feature extraction with artificial neural networks.After finishing my M.Sc.thesis I decided to pursue a Ph.D. in the fields of pattern recognition, machine learning, and artificial intelligence.When the Universiteit Maastricht offered me the opportunity to combine my research interests with my interest in Go, I did not hesitate.The research led to several conference papers, journal articles, and eventually this thesis.The research presented in this thesis has benefited from the help of many persons, whom I want to acknowledge here.First, I would like to thank my supervisor Jaap van den Herik.His tireless efforts to provide valuable feedback, even during his holidays, greatly improved the quality of the thesis.Next, many thanks to my daily advisor Jos Uiterwijk.Without the help of both of them this thesis would have never appeared.I would like to thank the members of the search and games group