Lemmatisation & morphological analysis of unedited Greek : do simple tasks need complex solutions?
- Author
- Colin Swaelens (UGent) , Ilse De Vos (UGent) and Els Lefever (UGent)
- Organization
- Project
- Abstract
- Fine-tuning transformer-based models for part-of-speech tagging of unedited Greek text has outperformed traditional systems. However, when applied to lemmatisation or morphological analysis, fine-tuning has not yet achieved competitive results. This paper explores various approaches to combine morphological features to both reduce label complexity and enhance multi-task training. Specifically, we group three nominal features into a single label, and combine the three most distinctive features of verbs into another unified label. These combined labels are used to fine-tune DBBERT, a BERT model pre-trained on both ancient and modern Greek. Additionally, we experiment with joint training -- both among these labels and in combination with POS tagging -- within a multi-task framework to improve performance by transferring parameters. To evaluate our models, we use a manually annotated gold standard from the Database of Byzantine Book Epigrams. Our results show a nearly 9 pp. improvement, demonstrating that multi-task learning is a promising approach for linguistic annotation in less standardised corpora.
- Keywords
- Natural Language Processing, Byzantine Greek, Lemmatisation, Morphological Analysis
Downloads
-
(...).pdf
- full text (Accepted manuscript)
- |
- UGent only
- |
- |
- 683.33 KB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-01JVC11RAW00KT3CBTP9YKHC0P
- MLA
- Swaelens, Colin, et al. “Lemmatisation & Morphological Analysis of Unedited Greek : Do Simple Tasks Need Complex Solutions?” Findings of the Association for Computational Linguistics : ACL 2025, edited by Wanxiang Che et al., Association for Computational Linguistics (ACL), 2025, pp. 7681–89.
- APA
- Swaelens, C., De Vos, I., & Lefever, E. (2025). Lemmatisation & morphological analysis of unedited Greek : do simple tasks need complex solutions? In W. Che, J. Nabende, E. Shutova, & M. T. Pilehvar (Eds.), Findings of the Association for Computational Linguistics : ACL 2025 (pp. 7681–7689). Association for Computational Linguistics (ACL).
- Chicago author-date
- Swaelens, Colin, Ilse De Vos, and Els Lefever. 2025. “Lemmatisation & Morphological Analysis of Unedited Greek : Do Simple Tasks Need Complex Solutions?” In Findings of the Association for Computational Linguistics : ACL 2025, edited by Wanxiang Che, Joyce Nabende, Ekaterina Shutova, and Mohammad Taher Pilehvar, 7681–89. Association for Computational Linguistics (ACL).
- Chicago author-date (all authors)
- Swaelens, Colin, Ilse De Vos, and Els Lefever. 2025. “Lemmatisation & Morphological Analysis of Unedited Greek : Do Simple Tasks Need Complex Solutions?” In Findings of the Association for Computational Linguistics : ACL 2025, ed by. Wanxiang Che, Joyce Nabende, Ekaterina Shutova, and Mohammad Taher Pilehvar, 7681–7689. Association for Computational Linguistics (ACL).
- Vancouver
- 1.Swaelens C, De Vos I, Lefever E. Lemmatisation & morphological analysis of unedited Greek : do simple tasks need complex solutions? In: Che W, Nabende J, Shutova E, Pilehvar MT, editors. Findings of the Association for Computational Linguistics : ACL 2025. Association for Computational Linguistics (ACL); 2025. p. 7681–9.
- IEEE
- [1]C. Swaelens, I. De Vos, and E. Lefever, “Lemmatisation & morphological analysis of unedited Greek : do simple tasks need complex solutions?,” in Findings of the Association for Computational Linguistics : ACL 2025, Vienna, Austria, 2025, pp. 7681–7689.
@inproceedings{01JVC11RAW00KT3CBTP9YKHC0P,
abstract = {{Fine-tuning transformer-based models for part-of-speech tagging of unedited Greek text has outperformed traditional systems. However, when applied to lemmatisation or morphological analysis, fine-tuning has not yet achieved competitive results. This paper explores various approaches to combine morphological features to both reduce label complexity and enhance multi-task training. Specifically, we group three nominal features into a single label, and combine the three most distinctive features of verbs into another unified label. These combined labels are used to fine-tune DBBERT, a BERT model pre-trained on both ancient and modern Greek. Additionally, we experiment with joint training -- both among these labels and in combination with POS tagging -- within a multi-task framework to improve performance by transferring parameters. To evaluate our models, we use a manually annotated gold standard from the Database of Byzantine Book Epigrams. Our results show a nearly 9 pp. improvement, demonstrating that multi-task learning is a promising approach for linguistic annotation in less standardised corpora.}},
author = {{Swaelens, Colin and De Vos, Ilse and Lefever, Els}},
booktitle = {{Findings of the Association for Computational Linguistics : ACL 2025}},
editor = {{Che, Wanxiang and Nabende, Joyce and Shutova, Ekaterina and Pilehvar, Mohammad Taher}},
isbn = {{9798891762565}},
keywords = {{Natural Language Processing,Byzantine Greek,Lemmatisation,Morphological Analysis}},
language = {{eng}},
location = {{Vienna, Austria}},
pages = {{7681--7689}},
publisher = {{Association for Computational Linguistics (ACL)}},
title = {{Lemmatisation & morphological analysis of unedited Greek : do simple tasks need complex solutions?}},
url = {{https://aclanthology.org/2025.findings-acl.399/}},
year = {{2025}},
}