{"id":2000810,"updated":"2025-02-27T01:51:46.604193+00:00","links":{},"created":"2025-02-21T01:44:04.985726+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:02000810","sets":["1164:1867:11897:1740054267812"]},"path":["1740054267812"],"owner":"80578","recid":"2000810","title":["大規模言語モデルのサービングにおける投機的デコーディングの動的最適化手法の提案"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2025-02-27"},"_buckets":{"deposit":"8995672a-9bb5-4d29-aeb2-c77ba604a096"},"_deposit":{"id":"2000810","pid":{"type":"depid","value":"2000810","revision_id":0},"owners":[80578],"status":"published","created_by":80578},"item_title":"大規模言語モデルのサービングにおける投機的デコーディングの動的最適化手法の提案","author_link":[],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"大規模言語モデルのサービングにおける投機的デコーディングの動的最適化手法の提案","subitem_title_language":"ja"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"リアルタイムシステム・性能最適化","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2025-02-27","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京大学"},{"subitem_text_value":"東京大学"}]},"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/2000810/files/IPSJ-OS25166004.pdf","label":"IPSJ-OS25166004.pdf"},"date":[{"dateType":"Available","dateValue":"2027-02-27"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-OS25166004.pdf","filesize":[{"value":"1.1 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":"11"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"d787f8f8-aded-40a1-abf6-a7c2610ec1ab","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":"薄井,真之"}]},{"creatorNames":[{"creatorName":"高前田,伸也"}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10444176","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":"2188-8795","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"投機的デコーディングを大規模言語モデルのサービング（デプロイメント）に適用するため、候補トークン数を動的に決定する方法と、リクエストを跨いでより見込みのあるトークンを（候補トークンとして）選択する方法を提案する。サービングシステム（サーバ）を実装することで評価を行い、提案手法がレイテンシを改善することを示す。","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"8","bibliographic_titles":[{"bibliographic_title":"研究報告システムソフトウェアとオペレーティング・システム（OS）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2025-02-27","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"4","bibliographicVolumeNumber":"2025-OS-166"}]},"relation_version_is_last":true,"weko_creator_id":"80578"}}