Development and Analysis of Shogi program with play style
| Project/Area Number |
22500123
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | The University of Electro-Communications |
|---|
Principal Investigator |
ITO Takeshi 電気通信大学, 大学院・情報理工学研究科, 助教 (40262373)
|
|---|
| Project Period (FY) |
2010 – 2012
|
| Project Status |
Completed (Fiscal Year 2012)
|
| Budget Amount *help |
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2012: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2011: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2010: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
|
|---|
| Keywords | プレイスタイル / 棋風 / ゲーム情報学 / 模倣AI / 機械学習 / コンピュータ将棋 / プレースタイル / 評価関数機械学習 / 主観的評価 |
|---|
| Research Abstract |
In this research, I proposed the following two approaches as method imitating play styles. I proposed a learning algorithm which overemphasizes the play style of a specific professional shogi player's by using the machine learning of Bonanza. By unbalanced learning of the play records of some top professional shogi players with features, I showed that the play styles of the players could be copied to some extent.I paid my attention to general play style of shogi; "offensive vs. defensive", "heavy vs. light", "linear vs. non-linear" and so on. I tried determination of the feature element which forms the play style by analyzing statistically the game records of a player which is reflecting each play style deeply.
|
Report
(4 results)
Research Products
(22 results)
