{"id":241580,"updated":"2025-01-19T07:36:52.729901+00:00","links":{},"created":"2025-01-19T01:46:16.141945+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00241580","sets":["1164:4179:11560:11869"]},"path":["11869"],"owner":"44499","recid":"241580","title":["Hybrid H3-Conformerを用いた頑健な長時間音声認識"],"pubdate":{"attribute_name":"公開日","attribute_value":"2024-12-05"},"_buckets":{"deposit":"95f671d2-50b0-448e-97fe-ad22a8a071dc"},"_deposit":{"id":"241580","pid":{"type":"depid","value":"241580","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"Hybrid H3-Conformerを用いた頑健な長時間音声認識","author_link":["665341","665343","665342"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Hybrid H3-Conformerを用いた頑健な長時間音声認識"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"ポスターセッション","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2024-12-05","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"京都大学"},{"subitem_text_value":"京都大学"},{"subitem_text_value":"京都大学"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"KU","subitem_text_language":"en"},{"subitem_text_value":"KU","subitem_text_language":"en"},{"subitem_text_value":"KU","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/241580/files/IPSJ-NL24262005.pdf","label":"IPSJ-NL24262005.pdf"},"date":[{"dateType":"Available","dateValue":"2026-12-05"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-NL24262005.pdf","filesize":[{"value":"546.3 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":"2e597888-b94d-471b-b67f-4cdb8f9a43ec","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":"本多, 智揮"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"坂井, 信輔"}],"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":"近年，多くの音声認識タスクにおいて Conformer は非常に高い性能を達成している．一方で，Conformer のような自己注意機構をもつモデルは，長時間の音声入力に対しては処理時間が非常に長くなるだけでなく，認識精度も低下することが知られている．この問題を解決するため，本研究では，自己注意機構の代替として構造化状態空間モデルの一種である Hungry Hungry Hippos (H3) を導入することを提案する．H3 の入力に線形な計算量で長い入力系列を効果的にモデル化できる特長を活用することにより，提案モデルである H3-Conformer は，長時間音声の効率的かつ堅牢な認識を実現した．さらに，H3 と自己注意機構を組み合わせたハイブリッドモデルを提案し，H3 を上位層で，自己注意機構を下位層で使用することで，オンライン音声認識タスクの性能を大幅に向上させることを示した．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告自然言語処理（NL）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2024-12-05","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"5","bibliographicVolumeNumber":"2024-NL-262"}]},"relation_version_is_last":true,"weko_creator_id":"44499"}}