{"created":"2025-01-19T01:20:21.988815+00:00","updated":"2025-01-19T14:36:18.792573+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00220321","sets":["1164:1579:10818:11010"]},"path":["11010"],"owner":"44499","recid":"220321","title":["オンライン逐次学習のための軽量コンセプトドリフト検知手法"],"pubdate":{"attribute_name":"公開日","attribute_value":"2022-10-04"},"_buckets":{"deposit":"795d73f6-3163-4d52-930d-daca90f8935f"},"_deposit":{"id":"220321","pid":{"type":"depid","value":"220321","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"オンライン逐次学習のための軽量コンセプトドリフト検知手法","author_link":["576045","576047","576046","576048"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"オンライン逐次学習のための軽量コンセプトドリフト検知手法"},{"subitem_title":"A Lightweight Concept Drift Detection Method for Online Sequential Learning                                                    ","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"IoT・エッジAI","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2022-10-04","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"慶應義塾大学大学院理工学研究科"},{"subitem_text_value":"慶應義塾大学大学院理工学研究科 "}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Science and Technology, Keio University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Science and Technology, Keio 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/220321/files/IPSJ-ARC22250008.pdf","label":"IPSJ-ARC22250008.pdf"},"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-ARC22250008.pdf","filesize":[{"value":"1.0 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"0","billingrole":"16"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_login","version_id":"7e00c058-1a22-4e46-bbd2-48bac952a190","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2022 by the Institute of Electronics, Information and Communication Engineers This SIG report is only available to those in membership of the SIG."}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"山田, 赳也"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"松谷, 宏紀"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Takeya, Yamada","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Hiroki, Matsutani","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10096105","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":"2188-8574","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"機械学習モデルの学習と予測をクラウドで行うのではなく，エッジデバイス上で両方を行うオンデバイス学習の関心が高まってきている．実世界のデータストリームの分布は一定ではなく，時間と伴に変化することがあり，データの変化によりモデルの精度の低下を招くことがあるため，データの特性が変化するたびにモデルの再学習を行う必要がある．このようなデータの特性が時間と伴に変化することをコンセプトドリフトと呼ぶ．本論文では，オンライン逐次学習のための軽量コンセプトドリフト検知手法を提案する．具体的には，入力データにおける各ラベルの平均座標を保持し，1 サンプルごとに平均座標の更新を行い，平均座標の移動距離でコンセプトドリフトを検知する．データ自体を蓄積する必要がないため，メモリ使用量を削減できる．2 種類のデータセットを用いて提案手法を評価したところ，既存手法と比べて精度は 2% から 26% ほど低下したが，メモリ使用量を最大で約 96% 削減できた．また，他の提案手法と比較して最大で約 83% 処理時間を短縮できた．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告システム・アーキテクチャ（ARC）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2022-10-04","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"8","bibliographicVolumeNumber":"2022-ARC-250"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":220321,"links":{}}