Elamur Research⌕
☰
home / papers / 2026 / BRAGD: Constrained Multi-Label POS Tagging for Faroese

BRAGD: Constrained Multi-Label POS Tagging for Faroese

Summary

Faroese is a morphologically rich North Germanic language with about 50,000 speakers and little annotated data. Earlier taggers treated each full morphological analysis as a single atomic tag, so many rare combinations appeared only a few times in training. The authors instead use a multi-label design. Their revised tagset, BRAGD, breaks 651 compound tags into 73 binary feature dimensions (word class, gender, number, case and others), so that frequent features can pool evidence across word classes.

Native Faroese linguists rebuilt the tagset around Faroese grammar rather than inherited Icelandic conventions. They also corrected the Sosialurin newspaper corpus, producing Sosialurin-BRAGD, and added a new 500-sentence out-of-domain test set. ScandiBERT is fine-tuned with a loss that hides features that cannot apply to a given word class, which also guarantees well-formed predictions at inference.

The tagger reaches 97.5% accuracy, clearly above ABLTagger evaluated on the same data, and it holds up better on newer text. The released models, data and code give Faroese a locally runnable, community-controlled tagger for parsing, proofing, research and teaching.

Key points

  • Accuracy gain: the constrained multi-label ScandiBERT tagger scores 97.5% overall. ABLTagger, the earlier bi-LSTM approach (Hafsteinsson, 2020), scores 89.9% when retrained on the revised corpus, a gap of 7.6 percentage points. ABLTagger had originally reported 91.4% on its older 371-tag scheme.
  • Corpus revision: Sosialurin-BRAGD has 6,099 sentences and 117,322 tokens. Of these, 27,426 tags were corrected or changed during the revision.
  • New tagset categories: BRAGD defines 15 word classes and adds distinctions that earlier Faroese schemes lacked:
    • a separate article category
    • supine verb forms
    • voice split from mood, including mediopassive
    • absolute superlative
    • explicit "no-feature" markers
  • Out-of-domain evaluation: the test set contains 500 sentences of contemporary 2019–2022 BLARK texts. On it, the constrained model significantly beats an unconstrained multi-label variant (p = 0.010, corrected Wilcoxon test).
  • Release and limitations: models, code and data are published at github.com/Maltoknidepilin/BRAGD. The authors note several open issues:
    • no full ablation study
    • weak results on classes with fewer than 50 training examples
    • prepositions are still grouped under adverbs

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Topics: Part-of-speech tagging

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/4ibuz9yubt3c

@inproceedings{simonsen2026bragd,
  title = {{BRAGD: Constrained Multi-Label POS Tagging for Faroese}},
  author = {Simonsen, Annika and Scalvini, Barbara and Johannesen, Uni and Debess, Iben Nyholm and Einarsson, Hafsteinn and Snæbjarnarson, Vésteinn},
  booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference},
  year = {2026},
  doi = {10.63317/4ibuz9yubt3c},
  url = {https://doi.org/10.63317/4ibuz9yubt3c}
}

References

41 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Yvonne Adesam and Aleksandrs Berdicevskis. 2021. Part-of-speech tagging of Swedish texts in the neural era. In Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa), pages 200–209, Reykjavik, Iceland (Online). Linköping University Electronic Press, Sweden. find
  2. Steven Bird, Ewan Klein, and Edward Loper. 2009. Natural Language Processing with Python. O’Reilly Media, Sebastopol, CA. find
  3. Terra Blevins, Hila Gonen, and Luke Zettlemoyer. 2023. Prompting language models for linguistic structure. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6649– 6663, Toronto, Canada. Association for Computational Linguistics. find
  4. Eibe Bouckaert, Remco R.and Frank. 2004. Evaluating the Replicability of Significance Tests for Comparing Learning Algorithms. In Advances in Knowledge Discovery and Data Mining, pages 3–12, Berlin, Heidelberg. Springer Berlin Heidelberg. find
  5. Thorsten Brants. 2000. TnT – A Statistical Part-of-Speech Tagger. In Sixth Applied Natural Language Processing Conference, pages 224–231, Seattle, Washington, USA. Association for Computational Linguistics. ↗ our card
  6. Grzegorz Chrupała, Georgiana Dinu, and Josef van Genabith. 2008. Learning Morphology with Morfette. In Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC’08), Marrakech, Morocco. European Language Resources Association (ELRA). find our card
  7. Ryan Cotterell and Georg Heigold. 2017. Cross-lingual Character-Level Neural Morphological Tagging. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 748–759, Copenhagen, Denmark. Association for Computational Linguistics. find
  8. Marie-Catherine de Marneffe, Christopher D. Manning, Joakim Nivre, and Daniel Zeman. 2021. Universal Dependencies. Computational Linguistics, 47(2):255–308. find our card
  9. Janez Demšar. 2006. Statistical Comparisons of Classifiers over Multiple Data Sets. Journal of Machine Learning Research, 7(1):1–30. find
  10. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 4171–4186, Minneapolis, Minnesota. Association for Computational Linguistics. ↗
  11. Hinrik Hafsteinsson. 2020. A Faroese part-of-speech tagger built with Icelandic methods. Data preparation, training and evaluation. Master’s thesis, The University of Iceland, September. find
  12. Hinrik Hafsteinsson and Anton Karl Ingason. 2020. Developing a Faroese PoS-tagging solution using Icelandic methods. In Proceedings of the 17th International Conference on Natural Language Processing (ICON), pages 481–490, Indian Institute of Technology Patna, Patna, India. NLP Association of India (NLPAI). find
  13. Zakaris Svabo Hansen, Heini Justinussen, and Mortan Ólason. 2004. Marking av teldutøkum tekstsavni [Tagging of a digital text corpus]. find
  14. Per E. Kummervold, Javier De la Rosa, Freddy Wetjen, and Svein Arne Brygfjeld. 2021. Operationalizing a National Digital Library: The Case for a Norwegian Transformer Model. In Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa), pages 20–29, Reykjavik, Iceland (Online). Linköping University Electronic Press, Sweden. ↗
  15. Andrey Kutuzov, Jeremy Barnes, Erik Velldal, Lilja Øvrelid, and Stephan Oepen. 2021. Large-scale contextualised language modelling for Norwegian. In Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa), pages 30–40, Reykjavik, Iceland (Online). Linköping University Electronic Press, Sweden. ↗
  16. John D. Lafferty, Andrew McCallum, and Fernando C. N. Pereira. 2001. Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data. In Proceedings of the Eighteenth International Conference on Machine Learning (ICML), pages 282–289. Morgan Kaufmann. find
  17. Hrafn Loftsson, Ida Kramarczyk, Sigrún Helgadóttir, and Eiríkur Rögnvaldsson. 2009. Improving the PoS tagging accuracy of Icelandic text. In Proceedings of the 17th Nordic Conference of Computational Linguistics (NODALIDA 2009), pages 103–110, Odense, Denmark. Northern European Association for Language Technology (NEALT). find
  18. Martin Malmsten, Love Börjeson, and Chris Haffenden. 2020. Playing with Words at the National Library of Sweden – Making a Swedish BERT. arXiv preprint arXiv:2007.01658. ↗ our card
  19. Beáta Megyesi. 2001. Comparing Data-Driven Learning Algorithms for PoS Tagging of Swedish. In Proceedings of the 2001 Conference on Empirical Methods in Natural Language Processing. find
  20. Thomas Müller, Helmut Schmid, and Hinrich Schütze. 2013. Efficient Higher-Order CRFs for Morphological Tagging. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 322–332, Seattle, Washington, USA. Association for Computational Linguistics. find
  21. Thomas Müller and Hinrich Schuetze. 2015. Robust Morphological Tagging with Word Representations. In Proceedings of the 2015 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 526–536, Denver, Colorado. Association for Computational Linguistics. find
  22. Claude Nadeau and Yoshua Bengio. 1999. Inference for the Generalization Error. In Advances in Neural Information Processing Systems, volume 12. MIT Press. find
  23. Graham Neubig, Chris Dyer, Yoav Goldberg, Austin Matthews, Waleed Ammar, Antonios Anastasopoulos, Miguel Ballesteros, David Chiang, Daniel Clothiaux, Trevor Cohn, Kevin Duh, Manaal Faruqui, Cynthia Gan, Dan Garrette, Yangfeng Ji, Lingpeng Kong, Adhiguna Kuncoro, Gaurav Kumar, Chaitanya Malaviya, Paul Michel, Yusuke Oda, Matthew Richardson, Naomi Saphra, Swabha Swayamdipta, and Pengcheng Yin. 2017. DyNet: The Dynamic Neural Network Toolkit. ↗ our card
  24. Robert Östling and Jörg Tiedemann. 2017. Continuous multilinguality with language vectors. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers, pages 644–649, Valencia, Spain. Association for Computational Linguistics. find
  25. Hjalmar P. Petersen, Christer Lindqvist, Jógvan í Lon Jacobsen, Zakaris Svabo Hansen, and Sissal M. Rasmussen, editors. 2025. Føroysk Málsøga 1. Nám, Tórshavn. find
  26. Slav Petrov, Dipanjan Das, and Ryan McDonald. 2012. A universal part-of-speech tagset. In Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC’12), pages 2089–2096, Istanbul, Turkey. European Language Resources Association (ELRA). find
  27. Jörgen Pind, Friðrik Magnússon, and Stefán Briem. 1991. Íslensk orðtíðnibók [The Icelandic Frequency Dictionary]. The Institute of Lexicography, University of Iceland, Reykjavík, Iceland. find
  28. Barbara Plank, Anders Søgaard, and Yoav Goldberg. 2016. Multilingual Part-of-Speech Tagging with Bidirectional Long Short-Term Memory Models and Auxiliary Loss. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 412–418, Berlin, Germany. Association for Computational Linguistics. find our card
  29. Miikka Silfverberg and Senka Drobac. 2018. Sub-label dependencies for Neural Morphological Tagging – The Joint Submission of University of Colorado and University of Helsinki for Var-Dial 2018. In Proceedings of the Fifth Workshop on NLP for Similar Languages, Varieties and Dialects (VarDial 2018), pages 37–45, Santa Fe, New Mexico, USA. Association for Computational Linguistics. find
  30. Annika Simonsen, Sandra Saxov Lamhauge, Iben Nyholm Debess, and Peter Juel Henrichsen. 2022. Creating a Basic Language Resource Kit for Faroese. In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 4637–4643, Marseille, France. European Language Resources Association. find
  31. Vésteinn Snæbjarnarson, Haukur Barri Símonarson, Pétur Orri Ragnarsson, Svanhvít Lilja Ingólfsdóttir, Haukur Jónsson, Vilhjalmur Thorsteinsson, and Hafsteinn Einarsson. 2022. A Warm Start and a Clean Crawled Corpus - A Recipe for Good Language Models. In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 4356–4366, Marseille, France. European Language Resources Association. find
  32. Vésteinn Snæbjarnarson, Annika Simonsen, Goran Glavaš, and Ivan Vulić. 2023. Transfer to a Low-Resource Language via Close Relatives: The Case Study on Faroese. In Proceedings of the 24th Nordic Conference on Computational Linguistics (NoDaLiDa), pages 728–737, Tórshavn, Faroe Islands. University of Tartu Library. ↗
  33. Statistics Faroe Islands. 2025. Population. Statistics Faroe Islands. Accessed December 2024. find
  34. Steinþór Steingrímsson, Örvar Kárason, and Hrafn Loftsson. 2019. Augmenting a BiLSTM tagger with a morphological lexicon and a lexical category identification step. In Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2019), pages 1161–1168, Varna, Bulgaria. INCOMA Ltd. find
  35. Francis Tyers, Mariya Sheyanova, Aleksandra Martynova, Pavel Stepachev, and Konstantin Vinogorodskiy. 2018. Multi-source synthetic treebank creation for improved cross-lingual dependency parsing. In Proceedings of the Second Workshop on Universal Dependencies (UDW 2018), pages 144–150, Brussels, Belgium. Association for Computational Linguistics. find
  36. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Ł ukasz Kaiser, and Illia Polosukhin. 2017. Attention is All you Need. In Advances in Neural Information Processing Systems, volume 30, pages 5998–6008. Curran Associates, Inc. 12. find
  37. Language Resource References Paulivar Andreasen and Árni Dahl. 1997. Mállæra. Føroya Skúlabókagrunnur, Tórshavn. find
  38. Certainly AI. 2020. Nordic BERT Models. Danish, Norwegian, and Swedish BERT models trained from scratch. find
  39. Jóhan Hendrik Winther Poulsen, Marjun Simonsen, Jógvan í Lon Jacobsen, Anfinnur Johansen, Zakaris Svabo Hansen, and Ragnar Sigrunarson. 1998. Føroysk orðabók. Orðabókagrunnurin. find
  40. Rúnarsson, Kristján and Jákupsson, Heðin and Hansen, Zakarias Svabo. 2023. góður. Føroyski bendingargrunnurin [Faroese Morphological Database]. Accessed: 2025-10-20. find
  41. Höskuldur Thráinsson, Hjalmar Páll Petersen, Jógvan í Lon Jacobsen, and Zakaris Svabo Hansen. 2012. Faroese. An overview and reference grammar, 3 edition. Faroe University Press/Linguistic Institute of Iceland, Tórshavn/Reykjavík. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/simonsen-2026-bragd-constrained-multi-label-pos/