{"updated":"2025-01-19T07:38:49.405662+00:00","links":{},"created":"2025-01-19T01:46:07.245102+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00241488","sets":["1164:2735:11468:11810"]},"path":["11810"],"owner":"44499","recid":"241488","title":["部分ドロップアウト法を用いたBayesian Neural Networkの改良"],"pubdate":{"attribute_name":"公開日","attribute_value":"2024-12-02"},"_buckets":{"deposit":"d0b2924e-323b-4984-bae3-2e495e924b66"},"_deposit":{"id":"241488","pid":{"type":"depid","value":"241488","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"部分ドロップアウト法を用いたBayesian Neural Networkの改良","author_link":["664910","664911"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"部分ドロップアウト法を用いたBayesian Neural Networkの改良"},{"subitem_title":"Improvement of Bayesian Neural Network Using Partial Dropout","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"2024-12-02","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"名古屋大学大学院情報学研究科"},{"subitem_text_value":"名古屋大学大学院情報学研究科"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Informatics, Nagoya University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Informatics, 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/241488/files/IPSJ-MPS24151008.pdf","label":"IPSJ-MPS24151008.pdf"},"date":[{"dateType":"Available","dateValue":"2026-12-02"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-MPS24151008.pdf","filesize":[{"value":"1.4 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":"a72e9888-baef-4159-a4b6-8159d49e5565","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2024 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"小松, 優治"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"北, 栄輔"}],"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_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8833","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"モンテカルロドロップアウト法は，ニューラルネットワーク (NN) で構成した予測モデルに対して，推論時にドロップアウトを用いたサンプリングを行うことでベイズ推定を行う．モンテカルロドロップアウト法は，実装が比較的容易であるが，評価のために多数のサンプリングを行う必要があり，計算コストが大きい．そこで，本研究では，モンテカルロドロップアウト法に基づく BNN の計算コストを改善する方法について述べる．提案手法では，モンテカルロドロップアウト法を用いてサンプリングするノードの範囲を限定することで計算コストを小さくする．これを部分モンテカルロドロップアウト法と呼ぶことにする．多層パーセプトロンモデルによる回帰問題に適用した結果，部分モンテカルロドロップアウト法を用いることで計算コストが下げるとともに，計算精度を向上する可能性があることがわかった．","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"The Monte Carlo dropout method perfroms Bays estimation by sampling evaluation using dropouts at the time of inference for the traditional Neural Network (NN) model. It is relatively easy to implement, but necessary to Perform a number of sampling for evaluation. This study describes the improvement of the computational cost of BNN based on the partial Monte Calro dropout method, in which the computational cost is reduced by limiting the number of the hyperparameters sampled by the Monte Calro dropout method. The effectiveness of the method is discussed in the multilayer perceptron model, and it was found that the use of the partial Monte Calro dropout method not only reduces the computational cost but also improves the computational accuracy.","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":"2024-12-02","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"8","bibliographicVolumeNumber":"2024-MPS-151"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":241488}