TU Darmstadt / ULB / TUbiblio

On the Role of Summary Content Units in Text Summarization Evaluation

Nawrath, Marcel ; Nowak, Agnieszka Wiktoria ; Ratz, Tristan ; Walenta, Danilo Constantin ; Opitz, Juri ; Ribeiro, Leonardo F. R. ; Sedoc, João ; Deutsch, Daniel ; Mille, Simon ; Liu, Yixin ; Gehrmann, Sebastian ; Zhang, Lining ; Mahamood, Saad ; Clinciu, Miruna ; Chandu, Khyathi ; Hou, Yufang (2024)
On the Role of Summary Content Units in Text Summarization Evaluation.
2024 Conference of the North American Chapter of the Association for Computational Linguistics. Mexico City, Mexico (16.06.2024 - 21.06.2024)
Konferenzveröffentlichung, Bibliographie

Dies ist die neueste Version dieses Eintrags.

Kurzbeschreibung (Abstract)

At the heart of the Pyramid evaluation method for text summarization lie human written summary content units (SCUs). These SCUs areconcise sentences that decompose a summary into small facts. Such SCUs can be used to judge the quality of a candidate summary, possibly partially automated via natural language inference (NLI) systems. Interestingly, with the aim to fully automate the Pyramid evaluation, Zhang and Bansal (2021) show that SCUs can be approximated by automatically generated semantic role triplets (STUs). However, several questions currently lack answers, in particular: i) Are there other ways of approximating SCUs that can offer advantages?ii) Under which conditions are SCUs (or their approximations) offering the most value? In this work, we examine two novel strategiesto approximate SCUs: generating SCU approximations from AMR meaning representations (SMUs) and from large language models (SGUs), respectively. We find that while STUs and SMUs are competitive, the best approximation quality is achieved by SGUs. We also show through a simple sentence-decomposition baseline (SSUs) that SCUs (and their approximations) offer the most value when rankingshort summaries, but may not help as much when ranking systems or longer summaries.

Typ des Eintrags: Konferenzveröffentlichung
Erschienen: 2024
Autor(en): Nawrath, Marcel ; Nowak, Agnieszka Wiktoria ; Ratz, Tristan ; Walenta, Danilo Constantin ; Opitz, Juri ; Ribeiro, Leonardo F. R. ; Sedoc, João ; Deutsch, Daniel ; Mille, Simon ; Liu, Yixin ; Gehrmann, Sebastian ; Zhang, Lining ; Mahamood, Saad ; Clinciu, Miruna ; Chandu, Khyathi ; Hou, Yufang
Art des Eintrags: Bibliographie
Titel: On the Role of Summary Content Units in Text Summarization Evaluation
Sprache: Englisch
Publikationsjahr: Juni 2024
Verlag: ACL
Buchtitel: The 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Proceedings of the Conference - Volume 2: Short Papers
Veranstaltungstitel: 2024 Conference of the North American Chapter of the Association for Computational Linguistics
Veranstaltungsort: Mexico City, Mexico
Veranstaltungsdatum: 16.06.2024 - 21.06.2024
URL / URN: https://aclanthology.org/2024.naacl-short.25/
Zugehörige Links:
Kurzbeschreibung (Abstract):

At the heart of the Pyramid evaluation method for text summarization lie human written summary content units (SCUs). These SCUs areconcise sentences that decompose a summary into small facts. Such SCUs can be used to judge the quality of a candidate summary, possibly partially automated via natural language inference (NLI) systems. Interestingly, with the aim to fully automate the Pyramid evaluation, Zhang and Bansal (2021) show that SCUs can be approximated by automatically generated semantic role triplets (STUs). However, several questions currently lack answers, in particular: i) Are there other ways of approximating SCUs that can offer advantages?ii) Under which conditions are SCUs (or their approximations) offering the most value? In this work, we examine two novel strategiesto approximate SCUs: generating SCU approximations from AMR meaning representations (SMUs) and from large language models (SGUs), respectively. We find that while STUs and SMUs are competitive, the best approximation quality is achieved by SGUs. We also show through a simple sentence-decomposition baseline (SSUs) that SCUs (and their approximations) offer the most value when rankingshort summaries, but may not help as much when ranking systems or longer summaries.

Fachbereich(e)/-gebiet(e): 01 Fachbereich Rechts- und Wirtschaftswissenschaften
01 Fachbereich Rechts- und Wirtschaftswissenschaften > Betriebswirtschaftliche Fachgebiete
01 Fachbereich Rechts- und Wirtschaftswissenschaften > Betriebswirtschaftliche Fachgebiete > Wirtschaftsinformatik
01 Fachbereich Rechts- und Wirtschaftswissenschaften > Betriebswirtschaftliche Fachgebiete > Fachgebiet Software Business & Information Management
20 Fachbereich Informatik
20 Fachbereich Informatik > Ubiquitäre Wissensverarbeitung
Hinterlegungsdatum: 11 Jul 2024 06:51
Letzte Änderung: 11 Jul 2024 06:51
PPN:
Export:
Suche nach Titel in: TUfind oder in Google

Verfügbare Versionen dieses Eintrags

Frage zum Eintrag Frage zum Eintrag

Optionen (nur für Redakteure)
Redaktionelle Details anzeigen Redaktionelle Details anzeigen