{"created":"2025-02-18T06:58:42.226365+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:02000477","sets":["1164:5159:1739855613862:1739855711540"]},"path":["1739855711540"],"owner":"80578","recid":"2000477","title":["Winning Tickets in Large Pretrained Speech Models: Exploring the Lottery Ticket Hypothesis in XLS-R"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2025-02-23"},"_buckets":{"deposit":"112b8d32-cb1c-4a40-b6dc-4ceb160cfa36"},"_deposit":{"id":"2000477","pid":{"type":"depid","value":"2000477","revision_id":0},"owners":[80578],"status":"published","created_by":80578},"item_title":"Winning Tickets in Large Pretrained Speech Models: Exploring the Lottery Ticket Hypothesis in XLS-R","author_link":[],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Winning Tickets in Large Pretrained Speech Models: Exploring the Lottery Ticket Hypothesis in XLS-R","subitem_title_language":"ja"},{"subitem_title":"Winning Tickets in Large Pretrained Speech Models: Exploring the Lottery Ticket Hypothesis in XLS-R","subitem_title_language":"en"}]},"item_type_id":"4","publish_date":"2025-02-23","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Informatics, Kyoto University"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"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/2000477/files/IPSJ-SLP25155004.pdf","label":"IPSJ-SLP25155004.pdf"},"date":[{"dateType":"Available","dateValue":"2027-02-23"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SLP25155004.pdf","filesize":[{"value":"788.0 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":"22"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"c718355d-26f6-4470-9ecd-da560ce6773b","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2025 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Jaeyoung,Lee"}]},{"creatorNames":[{"creatorName":"Tatsuya,Kawahara"}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Jaeyoung Lee","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"Tatsuya Kawahara","creatorNameLang":"en"}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10442647","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":"2188-8663","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"Large pretrained ASR models achieve state-of-the-art performance but are computationally expensive. The Lottery Ticket Hypothesis (LTH) hypothesizes that there exist sparse subnetworks, or “winning tickets,” that can match the performance of the full model. This study applies LTH to a large pretrained ASR model, namely XLS-R, demonstrating that winning tickets exist at up to 60% sparsity while maintaining accuracy. Using a subset of Common Voice covering 90 languages, we find that moderate pruning (80%-51% of the model size) enhances generalization, consistent with prior LTH findings. Our results confirm LTH's applicability to large ASR models, opening new avenues for efficient and scalable speech recognition.","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"Large pretrained ASR models achieve state-of-the-art performance but are computationally expensive. The Lottery Ticket Hypothesis (LTH) hypothesizes that there exist sparse subnetworks, or “winning tickets,” that can match the performance of the full model. This study applies LTH to a large pretrained ASR model, namely XLS-R, demonstrating that winning tickets exist at up to 60% sparsity while maintaining accuracy. Using a subset of Common Voice covering 90 languages, we find that moderate pruning (80%-51% of the model size) enhances generalization, consistent with prior LTH findings. Our results confirm LTH's applicability to large ASR models, opening new avenues for efficient and scalable speech recognition.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"4","bibliographic_titles":[{"bibliographic_title":"研究報告音声言語情報処理（SLP）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2025-02-23","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"4","bibliographicVolumeNumber":"2025-SLP-155"}]},"relation_version_is_last":true,"weko_creator_id":"80578"},"id":2000477,"updated":"2025-02-18T06:58:47.173204+00:00","links":{}}