{"updated":"2025-02-21T02:12:54.750273+00:00","links":{},"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00240765","sets":["6164:6165:6462:11854"]},"path":["11854"],"owner":"11","recid":"240765","title":["大規模言語モデルを用いたドメインスクワッティング検出法"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2024-10-15"},"_buckets":{"deposit":"8bcad809-970c-429d-9bc5-6a0cef9c06fd"},"_deposit":{"id":"240765","pid":{"type":"depid","value":"240765","revision_id":0},"owners":[11],"status":"published","created_by":11},"item_title":"大規模言語モデルを用いたドメインスクワッティング検出法","author_link":["661140","661141","661142","661143","661144","661145"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"大規模言語モデルを用いたドメインスクワッティング検出法","subitem_title_language":"ja"},{"subitem_title":"Domain Squatting Detection Using Large Language Models","subitem_title_language":"en"}]},"item_type_id":"18","publish_date":"2024-10-15","item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_18_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"NTTセキュリティホールディングス株式会社"},{"subitem_text_value":"NTTセキュリティホールディングス株式会社"},{"subitem_text_value":"NTTセキュリティホールディングス株式会社"}]},"item_18_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"NTT Security Holdings Corporation","subitem_text_language":"en"},{"subitem_text_value":"NTT Security Holdings Corporation","subitem_text_language":"en"},{"subitem_text_value":"NTT Security Holdings Corporation","subitem_text_language":"en"}]},"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/240765/files/IPSJ-CSS2024019.pdf","label":"IPSJ-CSS2024019.pdf"},"date":[{"dateType":"Available","dateValue":"2026-10-15"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-CSS2024019.pdf","filesize":[{"value":"842.6 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":"30"},{"tax":["include_tax"],"price":"0","billingrole":"46"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"cd38116c-9454-462f-b63d-b6a3f1b09403","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2024 by the Information Processing Society of Japan"}]},"item_18_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"千葉, 大紀"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"中野, 弘樹"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"小出, 駿"}],"nameIdentifiers":[{}]}]},"item_18_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Daiki, Chiba","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Hiroki, Nakano","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Takashi, Koide","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_5794","resourcetype":"conference paper"}]},"item_18_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"ドメインスクワッティングは，正規ドメイン名に類似したドメイン名を登録し，ユーザを欺く攻撃手法であり，継続的かつ深刻な脅威となっている．本研究では，大規模言語モデル（LLM）を活用した新たなドメインスクワッティング検出システムDomainLynxを提案する．従来手法の多くは，認知度の高いドメイン名に対する事前定義パターンに依存しており，認知度の低いドメイン名を標的とするものや新たな攻撃の検出に課題があった．これに対しDomainLynxは，LLMの文脈理解能力を活用することで，幅広いドメイン名の保護と新種のスクワッティング技術の検出が可能である．本システムの特徴は，LLMのハルシネーション（誤った情報生成）を軽減する専用モジュールにある．これにより，検出の信頼性と文脈適応性が向上し，多様なソースから得られる大規模データの効率的かつ正確な分析が可能となった．評価実験では，1,649件の既知のスクワッティングドメインを含むデータセットを用いて，Llama-3-70Bモデルによる検出精度94.7%を達成した．さらに，1ヶ月間の実環境テストでは，209万件の新規登録ドメインから34,359件のスクワッティングドメインを検出し，従来手法と比較して2.5倍の検出性能を示した．","subitem_description_type":"Other"}]},"item_18_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"This study introduces DomainLynx, a system leveraging Large Language Models (LLMs) to detect domain squatting, a persistent cybersecurity threat. Unlike existing methods that rely on predefined patterns for popular domains, DomainLynx utilizes LLMs' contextual understanding to protect a wide range of domains and identify new squatting techniques. A key innovation is DomainLynx's specialized module for mitigating LLM hallucinations, enhancing detection reliability and adaptability. This enables efficient analysis of large-scale data from diverse sources. Evaluation using 1,649 known squatting domains showed 94.7% detection accuracy with the Llama-3-70B model. In a month-long real-world test, DomainLynx identified 34,359 squatting domains from 2.09 million new registrations, outperforming existing methods by 2.5 times.","subitem_description_type":"Other"}]},"item_18_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"144","bibliographic_titles":[{"bibliographic_title":"コンピュータセキュリティシンポジウム2024論文集"}],"bibliographicPageStart":"137","bibliographicIssueDates":{"bibliographicIssueDate":"2024-10-15","bibliographicIssueDateType":"Issued"}}]},"relation_version_is_last":true,"weko_creator_id":"11"},"id":240765,"created":"2025-01-19T01:45:08.416328+00:00"}