Unveiling Computer Chess Evolution: Can Machine Learning Detect Historical Trends?

Andrea Borghesi, Paolo Ciancarini, Angelo Di Iorio, Gianluca Moro

Abstract: Computer Chess and AI are deeply intertwined, addressing similar research issues and proposing solutions that intersect. Even from a historical viewpoint, these two areas are strongly connected. This article presents a historical overview of Computer Chess research segmented in four seasons. Our research goal is to report an experiment of using cutting-edge machine learning tools to extract keywords and topics from a large set of scientific articles on Computer Chess, aiming at singling out the characterizing differences among the seasons. Moreover, we investigate the relationships between topics across seasons and their evolution. Although the seasons can be identified by clear milestones, we observed a lack of distinct boundaries between their topics. Instead, some issues recur across different seasons, albeit adjusted to new contexts, tools, and technologies.