Deep Convolutional Neural Network, Minorization-Maximization Algorithm, and Monte Carlo Tree Search on Block Go

• Item type: Symposium(1)
• 公開日: 2016-10-28
• タイトル: Deep Convolutional Neural Network, Minorization-Maximization Algorithm, and Monte Carlo Tree Search on Block Go
• タイトル / 言語: en
• タイトル: Deep Convolutional Neural Network, Minorization-Maximization Algorithm, and Monte Carlo Tree Search on Block Go
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: Block Go, Monte Carlo Tree Search, Minorization-maximization algorithm, Deep Convolutional\nNeural Network, Machine Learning
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_5794
• 資源タイプ: conference paper
• 著者所属: Department of Computer Science and Information Engineering, National Dong Hwa University
• 著者所属: Department of Computer Science and Information Engineering, National Dong Hwa University
• 著者所属: Department of Computer Science and Information Engineering, National Dong Hwa University
• 著者所属: Department of Applied Mathematics, Chung Yuan Christian University
• 著者所属(英): en
• 著者所属(英): Department of Computer Science and Information Engineering, National Dong Hwa University
• 著者所属(英): en
• 著者所属(英): Department of Computer Science and Information Engineering, National Dong Hwa University
• 著者所属(英): en
• 著者所属(英): Department of Computer Science and Information Engineering, National Dong Hwa University
• 著者所属(英): en
• 著者所属(英): Department of Applied Mathematics, Chung Yuan Christian University
• 著者名: Shi-Jim, Yen; Keng, Wen Li; Chingnung, Lin; Jr-Chang, Chen
• 著者名(英): Shi-Jim, Yen; Keng, Wen Li; Chingnung, Lin; Jr-Chang, Chen
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: Block Go is similar to the game of Go. The game is introduced by Pro Zhang Xu at 2009. His purpose is to reduce the complexity of Go and let the game be suitable for children. The complexity of Block Go is around 10^45, which is between checker and Othello. In this paper, we will state the rule Go and analyses the complexity of Block. Then, the Block Go program is implemented with Monte Carlo Tree Search (MCTS), and minorization-maximization pattern database. We also apply Deep Convolutional Neural Network (DCNN) on Block Go. In the future, we will apply transfer learning to improve the DCNN of Block Go, based on the numerous Go game records.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: Block Go is similar to the game of Go. The game is introduced by Pro Zhang Xu at 2009. His purpose is to reduce the complexity of Go and let the game be suitable for children. The complexity of Block Go is around 10^45, which is between checker and Othello. In this paper, we will state the rule Go and analyses the complexity of Block. Then, the Block Go program is implemented with Monte Carlo Tree Search (MCTS), and minorization-maximization pattern database. We also apply Deep Convolutional Neural Network (DCNN) on Block Go. In the future, we will apply transfer learning to improve the DCNN of Block Go, based on the numerous Go game records.
• 書誌情報: ゲームプログラミングワークショップ2016論文集; 巻 2016, p. 69-72, 発行日 2016-10-28
• 出版者 / 言語: ja
• 出版者: 情報処理学会