{"updated":"2025-01-22T17:35:08.108374+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00029375","sets":["1164:2240:2278:2281"]},"path":["2281"],"owner":"1","recid":"29375","title":["自然勾配学習法の有効性"],"pubdate":{"attribute_name":"公開日","attribute_value":"2001-05-25"},"_buckets":{"deposit":"0e1f216a-2d03-424a-8131-d750ba501779"},"_deposit":{"id":"29375","pid":{"type":"depid","value":"29375","revision_id":0},"owners":[1],"status":"published","created_by":1},"item_title":"自然勾配学習法の有効性","author_link":["0","0"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"自然勾配学習法の有効性"},{"subitem_title":"Efficiency of Natural Gradient Learning","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"2001-05-25","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"名古屋大学工学部　現東大"},{"subitem_text_value":"名古屋大学工学研究科計算理工学専攻"},{"subitem_text_value":"名古屋大学工学研究科計算理工学専攻"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Department of Mathematical Engineering and Information Physics, Faculty of Engineering, The University of Tokyo","subitem_text_language":"en"},{"subitem_text_value":"Department of Computational Science and Engineering,Graduate School of Engineering,Nagoya University","subitem_text_language":"en"},{"subitem_text_value":"Department of Computational Science and Engineering,Graduate School of Engineering,Nagoya 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/29375/files/IPSJ-HPC01086003.pdf"},"date":[{"dateType":"Available","dateValue":"2003-05-25"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-HPC01086003.pdf","filesize":[{"value":"223.9 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":"14"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"b25fb050-bdd7-4ea1-ab9d-b06b64e8e2d1","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2001 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"田中, 研太郎"},{"creatorName":"杉原, 正顯"},{"creatorName":"須田, 礼二"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Tanaka, Kentarou","creatorNameLang":"en"},{"creatorName":"Sugihara, Masaaki","creatorNameLang":"en"},{"creatorName":"Suda, Reiji","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10463942","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_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"ニューラルネットの学習において、学習の停滞期(プラトー)が起きて、なかなか学習が進まないことがある。そのプラトーを避け、もっと速く学習する方法として、自然勾配学習法が甘利らによって考えられた。本論文では、この自然勾配学習法がうまくいかない場合があることを示し、その解決策として、普通の勾配学習法と自然勾配学習法を組み合わせることを提案し、数値実験で有効性を示す。","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"Natural gradient learning (NGL) was proposed by Amari\\cite{Natural Gradient Works Efficiently in Learning}.In this paper, we show that NGL does not work well in some cases, and introduce combination of ordinal gradient learning (OGL) and NGL to solve the problem.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"17","bibliographic_titles":[{"bibliographic_title":"情報処理学会研究報告ハイパフォーマンスコンピューティング（HPC）"}],"bibliographicPageStart":"13","bibliographicIssueDates":{"bibliographicIssueDate":"2001-05-25","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"49(2001-HPC-086)","bibliographicVolumeNumber":"2001"}]},"relation_version_is_last":true,"weko_creator_id":"1"},"created":"2025-01-18T22:59:14.511869+00:00","id":29375,"links":{}}