{"links":{},"id":2007470,"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:02007470","sets":["1164:5305:1771205017738:1771205084619"]},"path":["1771205084619"],"owner":"80578","recid":"2007470","title":["異種AIが更新する思考データベースを用いた将棋における性能改善法"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2026-02-23"},"_buckets":{"deposit":"a53b140d-7102-47cc-a055-bb0523ece7d8"},"_deposit":{"id":"2007470","pid":{"type":"depid","value":"2007470","revision_id":0},"owners":[80578],"status":"published","created_by":80578},"item_title":"異種AIが更新する思考データベースを用いた将棋における性能改善法","author_link":[],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"異種AIが更新する思考データベースを用いた将棋における性能改善法","subitem_title_language":"ja"},{"subitem_title":"A Performance Improvement Method for Shogi AI Using a Reasoning Database Updated by Heterogeneous AIs","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"2026-02-23","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京情報デザイン専門職大学"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Tokyo Information Design Professional University","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/2007470/files/IPSJ-GI26057003.pdf","label":"IPSJ-GI26057003.pdf"},"date":[{"dateType":"Available","dateValue":"2028-02-23"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-GI26057003.pdf","filesize":[{"value":"483.9 KB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"18"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"7c2aba6e-cc7b-4b8f-910b-76117593bc78","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2026 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"竹内,章"}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Akira Takeuchi","creatorNameLang":"en"}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AA11362144","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8736","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"思考中に得られる情報を異種AIで共有し活用することは，多様性や創造性を高める効果が期待できる．将棋AIにおける一例として，α-β探索における置換表の情報をモンテカルロ木探索において活用することによって，性能改善する方法を検証した．結果として，置換表の最善手に対してUCB方策の式を調整し訪問回数を増加させる補正方法により，勝率を向上することができた．また，探索した深さと評価値の情報を用いた条件により確率的にリーフノードとする方法を追加することで，更に改善できる見通しを得た．","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"Sharing and utilizing information obtained during the reasoning process among heterogeneous AIs is expected to enhance diversity and creativity. As a case study in shogi AI, we investigated a method for improving performance by leveraging information stored in the transposition table of α-β search within Monte Carlo tree search (MCTS). The results showed that adjusting the UCB policy formula to increase the visit count for the best move recorded in the transposition table led to an improvement in win rate. Furthermore, we obtained prospects for additional improvement by introducing a method that probabilistically treats a node as a leaf based on the searched depth and evaluation value.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"4","bibliographic_titles":[{"bibliographic_title":"研究報告ゲーム情報学（GI）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2026-02-23","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"3","bibliographicVolumeNumber":"2026-GI-57"}]},"relation_version_is_last":true,"weko_creator_id":"80578"},"created":"2026-02-17T06:51:14.219723+00:00","updated":"2026-02-17T06:51:18.628258+00:00"}