{"id":242316,"updated":"2025-01-19T07:21:13.258269+00:00","links":{},"created":"2025-01-19T01:47:27.093795+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00242316","sets":["934:10195:11872:11922"]},"path":["11922"],"owner":"44499","recid":"242316","title":["クエリに対応した事前の要約を伴う大規模言語モデルによる企業事業概要生成"],"pubdate":{"attribute_name":"公開日","attribute_value":"2025-01-15"},"_buckets":{"deposit":"bed43f23-a5ef-472e-b9da-87700d176703"},"_deposit":{"id":"242316","pid":{"type":"depid","value":"242316","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"クエリに対応した事前の要約を伴う大規模言語モデルによる企業事業概要生成","author_link":["668826","668825"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"クエリに対応した事前の要約を伴う大規模言語モデルによる企業事業概要生成"},{"subitem_title":"Generating Corporate Business Overviews Using Large Language Models with Query-Responsive Pre-Summaries","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"[特集号投稿論文] 生成AI, RAG, 要約, 金融テキスト","subitem_subject_scheme":"Other"}]},"item_type_id":"3","publish_date":"2025-01-15","item_3_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"株式会社ユーザベース"}]},"item_3_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Uzabase, Inc.","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/242316/files/IPSJ-TDP0601008.pdf","label":"IPSJ-TDP0601008.pdf"},"date":[{"dateType":"Available","dateValue":"2025-01-15"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-TDP0601008.pdf","filesize":[{"value":"1.1 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"0","billingrole":"5"},{"tax":["include_tax"],"price":"0","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"9e467c8a-6262-4977-8c7a-7e720f928925","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2025 by the Information Processing Society of Japan"}]},"item_3_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"田村, 光太郎"}],"nameIdentifiers":[{}]}]},"item_3_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Koutarou, Tamura","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_3_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AA12894091","subitem_source_identifier_type":"NCID"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_6501","resourcetype":"journal article"}]},"item_3_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2435-6484","subitem_source_identifier_type":"ISSN"}]},"item_3_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"企業の事業状態を知るためには，高度に専門的知識を有し，高頻度に配信される報道を整理・理解することが必要であるが，アナリストなどの人手での情報の取得や整理は大きなコストがかかる．そのため，情報の機械的でかつ瞬時の整理・構造化が求められていて，自然言語処理技術を中心にした情報抽出技術が金融関係のテキスト処理で盛んに研究されてきた．昨今では，大規模言語モデルにより事業状態を具体的に説明するテキストを生成することが可能となり，抽出からテキストを生成することの需要が高まっている．本研究では，有価証券報告書や決算情報など多様なテキストデータを対象に，検索拡張生成（RAG）技術を適用し，企業の事業概要の要約文生成を試みた．複数種のデータの形式を統一するために，与えられたクエリに基づいた事前要約を行うことで，生成結果の質を向上させた．","subitem_description_type":"Other"}]},"item_3_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"To understand the state of a company's business, specialized accounting knowledge and the ability to quickly judge news reports are necessary. Therefore, natural language processing technology has been utilized to understand semantic structures and extract information. Recently, the demand for generating texts that concretely describe the business state using large language models has increased. In this study, we applied Retrieval Augmented Generation (RAG) technology to various text data such as securities reports and financial information, and attempted to generate summary texts of a company's business overview. By performing pre-summaries based on the given queries to unify the formats of multiple types of data, we improved the quality of the generated results.","subitem_description_type":"Other"}]},"item_3_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"53","bibliographic_titles":[{"bibliographic_title":"情報処理学会論文誌デジタルプラクティス（TDP）"}],"bibliographicPageStart":"44","bibliographicIssueDates":{"bibliographicIssueDate":"2025-01-15","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"1","bibliographicVolumeNumber":"6"}]},"relation_version_is_last":true,"weko_creator_id":"44499"}}