{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00113274","sets":["1164:2735:7896:7897"]},"path":["7897"],"owner":"11","recid":"113274","title":["差分プライベート弱学習器の統合"],"pubdate":{"attribute_name":"公開日","attribute_value":"2015-02-24"},"_buckets":{"deposit":"ba0f8577-20fd-4a3e-8798-0551f6c6cd0f"},"_deposit":{"id":"113274","pid":{"type":"depid","value":"113274","revision_id":0},"owners":[11],"status":"published","created_by":11},"item_title":"差分プライベート弱学習器の統合","author_link":["37605","37610","37608","37609","37607","37612","37611","37606"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"差分プライベート弱学習器の統合"},{"subitem_title":"Aggregating Differentially Private Weak Learners","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"2015-02-24","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京大学"},{"subitem_text_value":"東京大学"},{"subitem_text_value":"東京大学"},{"subitem_text_value":"東京大学"}]},"item_4_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 Tokyo","subitem_text_language":"en"},{"subitem_text_value":"The University of Tokyo","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/113274/files/IPSJ-MPS15102001.pdf"},"date":[{"dateType":"Available","dateValue":"2017-02-24"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-MPS15102001.pdf","filesize":[{"value":"1.1 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"17"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"1206f5aa-2077-49b4-8182-c0ccab0dbf8c","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2015 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"南, 賢太郎"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"佐藤, 一誠"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"荒井, ひろみ"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"中川, 裕志"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Kentaro, Minami","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Issei, Sato","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Hiromi, Arai","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Hiroshi, Nakagawa","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10505667","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":"データが複数の組織にわたり分散して存在しているとき，それらを互いに共有することで，各組織におけるデータ解析の精度向上が期待できる．しかし，保護すべき個人情報がデータに含まれている場合には，異なる組織間での情報の交換は，プライバシ保護を考慮した上で行われなければならない．本研究では，差分プライバシをみたす弱学習器を互いに交換し，それらを統合する枠組みを提案する．これによって，複数の組織に分散したデータからの学習を，個人情報を保護しつつ効率的に行うことができる．また，特に学習タスクが二値分類である場合について計算機実験を行い，提案手法の性能を評価する．","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"When the dataset is distributed over a number of organizations, one can expect the improvement of data analysis by sharing the dataset each other. However, if the dataset consists of personal information, data sharing procedures must be performed under privacy-preserving constraints. Recently, differentially private algorithms for some statistical learning problems, such as empirical risk minimization, have been considered by several authors. In this work, we introduce a general framework for exponential weighting aggregation (EWA) of differentially private weak learners. This framework allows us to learn effectively from distributed dataset without leakage of personal information. Especially in the case of the binary classification problem, we evaluate the effectiveness of our approach on synthetic and real dataset.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告数理モデル化と問題解決（MPS）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2015-02-24","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"1","bibliographicVolumeNumber":"2015-MPS-102"}]},"relation_version_is_last":true,"weko_creator_id":"11"},"id":113274,"updated":"2025-01-20T19:42:05.818257+00:00","links":{},"created":"2025-01-18T23:54:53.393277+00:00"}