{"updated":"2025-01-21T13:34:43.321640+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00095813","sets":["6164:6165:6210:7301"]},"path":["7301"],"owner":"11","recid":"95813","title":["将棋の評価関数の学習に有用な局面の自動選択"],"pubdate":{"attribute_name":"公開日","attribute_value":"2013-11-01"},"_buckets":{"deposit":"3c31c87b-d357-4539-b3fd-47c9f024690f"},"_deposit":{"id":"95813","pid":{"type":"depid","value":"95813","revision_id":0},"owners":[11],"status":"published","created_by":11},"item_title":"将棋の評価関数の学習に有用な局面の自動選択","author_link":["0","0"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"将棋の評価関数の学習に有用な局面の自動選択"},{"subitem_title":"Automatic Selection of Useful Positions for Learning Shogi Evaluation Functions","subitem_title_language":"en"}]},"item_type_id":"18","publish_date":"2013-11-01","item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_18_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京大学工学部電子情報工学科"},{"subitem_text_value":"東京大学大学院工学系研究科"},{"subitem_text_value":"マンチェスター大学コンピュータ科学科"},{"subitem_text_value":"東京大学大学院工学系研究科"},{"subitem_text_value":"東京大学大学院工学系研究科"}]},"item_18_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"The University of Tokyo","subitem_text_language":"en"},{"subitem_text_value":"The University of Tokyo","subitem_text_language":"en"},{"subitem_text_value":"The University of Manchester","subitem_text_language":"en"},{"subitem_text_value":"The University of Tokyo","subitem_text_language":"en"},{"subitem_text_value":"The University of Tokyo","subitem_text_language":"en"}]},"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/95813/files/IPSJ-GPWS2013010.pdf"},"date":[{"dateType":"Available","dateValue":"2013-11-01"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-GPWS2013010.pdf","filesize":[{"value":"258.9 kB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"0","billingrole":"5"},{"tax":["include_tax"],"price":"0","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"61335630-b207-43bd-b00e-c2c41e7068d2","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2013 by the Information Processing Society of Japan"}]},"item_18_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"川上, 裕生"},{"creatorName":"浦, 晃"},{"creatorName":"三輪, 誠"},{"creatorName":"鶴岡, 慶雅"},{"creatorName":"近山, 隆"}],"nameIdentifiers":[{}]}]},"item_18_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Yusei, Kawakami","creatorNameLang":"en"},{"creatorName":"Akira, Ura","creatorNameLang":"en"},{"creatorName":"Makoto, Miwa","creatorNameLang":"en"},{"creatorName":"Yoshimasa, Tsuruoka","creatorNameLang":"en"},{"creatorName":"Takashi, Chikayama","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_5794","resourcetype":"conference paper"}]},"item_18_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"将棋プログラムの評価関数は大量の棋譜を利用した機械学習によって調整する。これにはプロ棋士\nの棋譜が用いられているが棋譜の数には限りがあり、新たに指し手の教師情報のついた局面を作成するに\nは大きなコストが必要となる。本稿では、教師情報を付けるコストを削減するために、能動学習を用いて\n学習に有効に働く局面を選択する手法を提案する。既存の棋譜を用いて提案手法の評価を行ったところ、\n将棋の評価関数の学習に有効な局面が存在し、その選択が可能であることを示した。","subitem_description_type":"Other"}]},"item_18_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"Today, the parameters of evaluation function of a shogi program are usually adjusted by ma-\nchine learning methods using many game records. This approach requires game records of professional Shogi\nplayers, but the number of such game records is limited and creating training data is costly. In this paper,\nto reduce the cost of creating training data, we propose an active learning-based method to select positions\nthat are particularly useful for learning. We evaluated the proposed method by using existing game records.\nExperimental results show that that there indeed exist such positions and that it is possible to select them\nautomatically.","subitem_description_type":"Other"}]},"item_18_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"72","bibliographic_titles":[{"bibliographic_title":"ゲームプログラミングワークショップ2013論文集"}],"bibliographicPageStart":"66","bibliographicIssueDates":{"bibliographicIssueDate":"2013-11-01","bibliographicIssueDateType":"Issued"}}]},"relation_version_is_last":true,"weko_creator_id":"11"},"created":"2025-01-18T23:42:48.775287+00:00","id":95813,"links":{}}