{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00227353","sets":["1164:5064:11199:11312"]},"path":["11312"],"owner":"44499","recid":"227353","title":["deepGTTM-IV: 深層学習に基づくタイムスパン木分析器"],"pubdate":{"attribute_name":"公開日","attribute_value":"2023-08-20"},"_buckets":{"deposit":"a9f8e542-2f37-4f02-888c-eaccffb1d8f7"},"_deposit":{"id":"227353","pid":{"type":"depid","value":"227353","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"deepGTTM-IV: 深層学習に基づくタイムスパン木分析器","author_link":["605479"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"deepGTTM-IV: 深層学習に基づくタイムスパン木分析器"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"深層学習","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2023-08-20","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"理化学研究所革新知能統合研究センター(AIP)"}]},"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/227353/files/IPSJ-MUS23138014.pdf","label":"IPSJ-MUS23138014.pdf"},"date":[{"dateType":"Available","dateValue":"2025-08-20"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-MUS23138014.pdf","filesize":[{"value":"3.2 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":"21"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"8a688620-c9af-4a08-b561-851459fb059d","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2023 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"浜中, 雅俊"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10438388","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":"2188-8752","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"本稿では,深層学習に基づく Generative Theory of Tonal Music(GTTM)のタイムスパン木分析器の構築について述べる.従来からタイムスパン木分析器の構築は試行されてきたが,多くの分析器では性能が非常に低く,比較的性能が良いものではパラメータを手動で調整する必要があった.これまでに我々はタイムスパン木の枝を 1 つずつ減らしていく逐次簡約法を提案し,Transformer モデルを用いて学習できることを確認した.しかし,逐次簡約法では,簡約された音符がどの音符に吸収されたかが明確でないためタイムスパン木を獲得することができなかった.そこで,逐次簡約法において学習時の符号化方法を改良し,どの音符がどの音符に簡約されるかを明示的に表すことにした.また,タイムスパン木を行列として表現することで,逐次簡約法を反復するタイムスパン木獲得アルゴリズムを提案する.学習に使ってない曲のタイムスパン木の分析を試みたところ,30 の曲のうち 29 曲で正しいタイムスパン木が得られた.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"9","bibliographic_titles":[{"bibliographic_title":"研究報告音楽情報科学(MUS)"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2023-08-20","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"14","bibliographicVolumeNumber":"2023-MUS-138"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":227353,"updated":"2025-01-19T12:12:03.334815+00:00","links":{},"created":"2025-01-19T01:26:37.201000+00:00"}