Computer Go Research Articles

Exploring the application of artificial intelligence techniques relating to AlphaGo in the diagnosis and treatment of traditional Chinese medicine

Go is not only regarded as the most complex board game, but also viewed as a huge challenge for artificial intelligence (AI). A computer Go program, AlphaGo, using artificial neural networks and tree search defeated one of the best professional human players in March 2016, achieving an unprecedented performance in the game of Go. This victory represented a major milestone in AI research. AI would profoundly affect all fields in the days to come. On the other hand, traditional Chinese medicine (TCM) is thought to be of similar cultural background, traditional concept and way of thinking to Go. Consequently AI techniques may be employed in TCM research. This article presented a brief overview of recent progress in AI, and explored the feasibility of applying the techniques concerning AlphaGo to the diagnosis and treatment of TCM. Key words: Artificial intelligence; AlphaGo; Diagnosis and treatment of traditional Chinese medicine

On Monte Carlo tree search and reinforcement learning

Fuelled by successes in Computer Go, Monte Carlo tree search (MCTS) has achieved wide-spread adoption within the games community. Its links to traditional reinforcement learning (RL) methods have b...

コンピュータ囲碁の進歩

コンピュータ囲碁の進歩

Review of deep reinforcement learning and discussions on the development of computer Go

Review of deep reinforcement learning and discussions on the development of computer Go

The 6th GPW Cup 9x9 and 13x13 Computer Go Tournaments

The 6th GPW Cup 9x9 and 13x13 Computer Go Tournaments

The Design and Development of a 9 by 9 GO Opening Game Knowledge Base System

This study used the hashing method to produce a 9 by 9 GO opening game knowledge base system, and conducted in-depth study to solve the various board situation symmetry problems. This system can analyze the winning rate of various moves in the opening game. If a sufficient number of chess manuals can be inputted into the system, it can effectively provide move choices in the computer GO opening game, and can be used for the references of GO players on opening games. In sum, according to this knowledge base system of this article, this study considers that computer GO may defeat human experts perfectly in the future.

Effect of Potential Model Pruning on Different-Sized Boards in Monte-Carlo GO

Monte-Carlo GO is a computer GO program that is sufficiently competent without the knowledge expressions of IGO. Although it is computationally intensive, the computational complexity can be reduced by properly pruning the IGO game tree. In this study, we achieved this by using a potential model based on the knowledge expressions of IGO. The potential model treats GO stones as potentials. A specific potential distribution on the GO board results from a unique arrangement of stones on the board. Pruning with the potential model categorizes legal moves into effective and ineffective moves in accordance with the potential threshold. In this study, certain pruning strategies based on potentials and potential gradients were experimentally evaluated. In particular, for different-sized boards, the effects of pruning strategies were evaluated in terms of their robustness. We successfully demonstrated pruning with a potential model to reduce the computational complexity of the game of GO as well as the robustness of this effect across different-sized boards.

Human vs. Computer Go Competition in FUZZ-IEEE 2011

Human vs. Computer Go Competition in FUZZ-IEEE 2011

The 4th GPW Cup 9x9 and 13x13 Computer-Go Tournaments

The 4th GPW Cup 9x9 and 13x13 Computer-Go Tournaments

The 2nd UEC Cup Computer-Go Tournament

The 2nd UEC Cup Computer-Go Tournament

Human vs. Computer-Go Competition

Human vs. Computer-Go Competition

Automatic pattern acquisition from game records in GO

Computer programs of GO are typically constructed using a knowledge-based approach with heuristics and pattern matching because of enormous complexities of the game. In this approach, quantity, quality, and consistency of patterns used in computer programs of GO to a large extent determine the strengths of the programs. This study presents an effective method to acquire automatically comprehensive GO patterns from large collections of game records. Statistical usages of the patterns ensure consistency and quality of the patterns, which in turn can help improve the strengths of computer GO programs. Additionally, statistical usages of patterns from different sources of game records clearly show subtle and significant discrepancies among various types of GO players, and clarify certain myths in the playing of GO.

MATHEMATICAL MORPHOLOGY APPLIED TO COMPUTER GO

This paper shows how mathematical morphological operators can be applied to computer go. On the one hand, mathematical morphology is a very powerful tool within image processing community. On the other hand, the Zobrist's model is well known within the computer go community for its "influence" recognition. We present a model, derived from the closing operator of mathematical morphology and from the Zobrist's model, which yields very good results for "territory" recognition. Moreover, we give efficient implementations of the dilation operator and territory recognition for computer go. This model was found when developing Indigo, our go playing program, and is now used with success in GnuGo, the go playing program of the Free Software Foundation.

A Report on the 4th Fost-Cup World-Open Computer-Go Championships

A Report on the 4th Fost-Cup World-Open Computer-Go Championships

The 3rd Fost-Cup World-Open Computer-Go Championship

The 3rd Fost-Cup World-Open Computer-Go Championship

Why should computers go parallel?

Why should computers go parallel?

Computers Go Home?

Computers Go Home?