{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00231270","sets":["1164:4179:11237:11430"]},"path":["11430"],"owner":"44499","recid":"231270","title":["調音属性に関する知識の埋め込みによるアイヌ語音声認識の改善"],"pubdate":{"attribute_name":"公開日","attribute_value":"2023-11-25"},"_buckets":{"deposit":"00f3c733-377f-41c4-8201-639609722677"},"_deposit":{"id":"231270","pid":{"type":"depid","value":"231270","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"調音属性に関する知識の埋め込みによるアイヌ語音声認識の改善","author_link":["624106","624105"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"調音属性に関する知識の埋め込みによるアイヌ語音声認識の改善"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"分野横断（1）","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2023-11-25","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"京都大学大学院情報学研究科"},{"subitem_text_value":"京都大学大学院情報学研究科"}]},"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/231270/files/IPSJ-NL23258012.pdf","label":"IPSJ-NL23258012.pdf"},"date":[{"dateType":"Available","dateValue":"2025-11-25"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-NL23258012.pdf","filesize":[{"value":"449.1 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":"23"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"80ec4289-fee6-453d-b3ac-2a43ad0a8df2","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":"李, 在詠"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"河原, 達也"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10115061","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":"2188-8779","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"本研究では，音素とその調音属性に関する知識を，低資源言語の自動音声認識の改善に活用する方法を提案する．まず，レイヤー初期化において音素と調音属性の関係の制約を埋め込むことで「調音属性予測レイヤー」を構築し，明示的な学習なしで調音属性を予測できるようにする．最終的な出力は，このレイヤーの出力と音響的特徴を組み合わせて推論する．この手法を大規模事前学習モデルに基づく End-to-End 音声認識に実装し，ファインチューニングを行う．アイヌ語の音声認識で評価を行い，ファインチューニングのデータが1時間しか利用できない場合に約 12% の相対的な改善を達成できた．これは，大規模事前学習モデルと組み合わせたときに音声学的な事前知識を取り入れるアプローチが有用であることを示している．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告自然言語処理（NL）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2023-11-25","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"12","bibliographicVolumeNumber":"2023-NL-258"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":231270,"updated":"2025-01-19T10:49:45.650462+00:00","links":{},"created":"2025-01-19T01:31:28.292177+00:00"}