{"links":{},"id":2007730,"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:02007730","sets":["1164:4179:1771495534039:1771495585049"]},"path":["1771495585049"],"owner":"80578","recid":"2007730","title":["多段階プロンプトを用いた推論過程の改善による大規模言語モデルの翻訳性能の向上"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2026-02-28"},"_buckets":{"deposit":"8814d3ec-215c-4fa1-839a-4c8b30bb829b"},"_deposit":{"id":"2007730","pid":{"type":"depid","value":"2007730","revision_id":0},"owners":[80578],"status":"published","created_by":80578},"item_title":"多段階プロンプトを用いた推論過程の改善による大規模言語モデルの翻訳性能の向上","author_link":[],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"多段階プロンプトを用いた推論過程の改善による大規模言語モデルの翻訳性能の向上","subitem_title_language":"ja"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"機械翻訳","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2026-02-28","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"東京科学大学"},{"subitem_text_value":"東京科学大学"},{"subitem_text_value":"東京科学大学"},{"subitem_text_value":"東京科学大学／産業技術総合研究所"},{"subitem_text_value":"東京科学大学／産業技術総合研究所／国立情報学研究所大規模言語モデル研究開発センター"}]},"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/2007730/files/IPSJ-NL26267004.pdf","label":"IPSJ-NL26267004.pdf"},"date":[{"dateType":"Available","dateValue":"2028-02-28"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-NL26267004.pdf","filesize":[{"value":"1.8 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":"23"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"8ae84e13-da20-4c60-bc1c-31cf9f6efb95","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2026 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"松下,直矢"}]},{"creatorNames":[{"creatorName":"Youmi,Ma"}]},{"creatorNames":[{"creatorName":"大葉,大輔"}]},{"creatorNames":[{"creatorName":"水木,栄"}]},{"creatorNames":[{"creatorName":"岡崎,直観"}]}]},"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":"最終回答の前に推論過程を展開する推論型の大規模言語モデルは，数学やコード生成で優れた性能を示すが，その推論過程の有効性はタスクに依存しており，時にはタスクに応じた設計が必要であることが示唆されている．本稿では，プロンプティングを用いてモデルの推論過程を制御することで，機械翻訳に有効な推論過程を体系的に探索する．具体的には，プロンプトの段数を増やす「縦方向」と，下訳の生成数を増やす「横方向」の2軸でプロンプトを設計した上で，推論コストのスケーリングに伴う翻訳性能の変化を検証した．実験の結果，縦方向・横方向ともに設計したプロンプトの有効性が確認され，横方向は高い翻訳性能を実現する一方，縦方向では推論コストに対する性能改善効率に優れることが示唆された．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"13","bibliographic_titles":[{"bibliographic_title":"研究報告自然言語処理（NL）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2026-02-28","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"4","bibliographicVolumeNumber":"2026-NL-267"}]},"relation_version_is_last":true,"weko_creator_id":"80578"},"created":"2026-02-19T10:20:20.573585+00:00","updated":"2026-02-19T10:20:24.315037+00:00"}