@inproceedings{424272cd1c0f45a4b59790466cac26fc,
title = "DanSumT5: Automatic Abstractive Summarization for Danish",
abstract = "Automatic abstractive text summarization is a challenging task in the field of natural language processing. This paper presents a model for domain-specific summarization for Danish news articles. DanSumT5 is an mT5 model fine-tuned on a cleaned subset of the DaNewsroom dataset comprising abstractive article-summary pairs. The resulting state-of-the-art model is evaluated both quantitatively and qualitatively, using ROUGE and BERTScore metrics, along with human rankings of the summaries. We find that although model refinements increase quantitative and qualitative performance, the model is still prone to factual errors. We discuss the limitations of current evaluation methods for automatic abstractive summarization and underline the need for improved metrics and transparency within the field. We suggest that future work should employ techniques for detecting and reducing errors in model output and methods for reference-less evaluation of summaries.",
author = "Sara Kolding and Katrine Nymann and Hansen, {Ida Bang} and Kenneth Enevoldsen and Kristensen-McLachlan, {Ross Deans}",
note = "Publisher Copyright: {\textcopyright} 2023 Association for Computational Linguistics.; 24th Nordic Conference on Computational Linguistics, NoDaLiDa 2023 ; Conference date: 22-05-2023 Through 24-05-2023",
year = "2023",
language = "English",
isbn = "978-99-1621-999-7",
series = "NEALT (Northern European Association of Language Technology) Proceedings Series",
pages = "248--264",
editor = "Tanel Alum{\"a}e and Mark Fishel",
booktitle = "Proceedings of the 24th Nordic Conference on Computational Linguistics (NoDaLiDa)",
publisher = "Tartu University Press",
}