Sudden-death prediction using Deep Convolutional Neural Network in Connect6

• Item type: Symposium(1)
• 公開日: 2022-11-04
• タイトル / 言語: en
• タイトル: Sudden-death prediction using Deep Convolutional Neural Network in Connect6
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: Connect6
• キーワード / 主題Scheme: Other
• キーワード / 主題: Deep Convolutional Neural Network
• キーワード / 主題Scheme: Other
• キーワード / 主題: Sudden-death positions
• 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(英): en
• 著者所属(英): School Of Information Science and Technology, Huizhou University, Huicheng District, Huizhou, Guangdong, China
• 著者所属(英): en
• 著者所属(英): Dept. Of Computer Science and Information Engineering, National Dong Hwa University, Taiwan
• 著者所属(英): en
• 著者所属(英): Dept. Of Computer Science and Information Engineering, National Dong Hwa University, Taiwan
• 著者名: Jung-Kuei, Yang; Shi-Jim, Yen; Yu-Yu, Yang
• 著者名(英): Jung-Kuei, Yang; Shi-Jim, Yen; Yu-Yu, Yang
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: This paper describes using Deep Convolutional Neural Network (DCNN) to predict the positions of sudden-death in Connect6. In sudden-death game, if one side cannot identify sudden-death positions, the other side will win the game at next move. Therefore, the prediction of sudden-death positions is of great significance for pruning the branch degree of the search tree. This study proposes many deep CNN model based on the features of Connect6 and trains it by lots of sudden-death positions established from our previous study. Then the best DCNN model is selected from the experimental results. The experimental results show that the depth of stacking multiple convolutional layers is the key influencing factor of deep CNN to predict sudden-death positions in Connect6. The results of this study can improve the search performance of Kavalan, which is an AI program we design to play Connect6 game.
• 書誌情報: ゲームプログラミングワークショップ2022論文集; 巻 2022, p. 243-246, 発行日 2022-11-04
• 出版者 / 言語: ja
• 出版者: 情報処理学会