Exploring GnuGo's Evaluation Function with a SVM

Abstract

While computers have defeated the best human players in many classic board games, progress in Go remains elusive. The large branching factor in the game makes traditional adversarial search intractable while the complex interaction of stones makes it difficult to assign a reliable evaluation function. This is why most existing programs rely on hand-tuned heuristics and pattern matching techniques. Yet none of these solutions perform better than an amateur player. Our work introduces a composite approach, aiming to integrate the strengths of the proved heuristic algorithms, the AI-based learning techniques, and the knowledge derived from expert games. Specifically, this paper presents an application of the Support Vector Machine (SVM) for training the GnuGo evaluation function.