Towards better language representation in Natural Language Processing
Abstract This paper introduces MultiGEC, a dataset for multilingual Grammatical Error Correction (GEC) in twelve European languages: Czech, English, Estonian, German, Greek, Icelandic, Italian, Latvian, Russian, Slovene, Swedish and Ukrainian. MultiGEC distinguishes itself from previous GEC datasets in that it covers several underrepresented languages, which we argue should be included in resources used to train models for Natural Language Processing tasks which, as GEC itself, have implications for Learner Corpus Research and Second Language Acquisition. Aside from multilingualism, the novelty of the MultiGEC dataset is that it consists of full texts — typically learner essays — rather than individual sentences, making it possible to train systems that take a broader context into account. The dataset was built for MultiGEC-2025, the first shared task in multilingual text-level GEC, but it remains accessible after its competitive phase, serving as a resource to train new error correction systems and perform cross-lingual GEC studies.
Authors
- Aleš Žagar (ORCID: https://orcid.org/0000-0002-4922-1547)
- Katrin Wisniewski (ORCID: https://orcid.org/0000-0003-4124-2543)
- Murathan Kurfalı (ORCID: https://orcid.org/0000-0002-7020-8275)
- Lionel Nicolas (ORCID: https://orcid.org/0000-0002-7703-9953)
- Roberts Darģis (ORCID: https://orcid.org/0000-0001-9375-6410)
- Kristjan Suluste
- Jennifer-Carmen Frey (ORCID: https://orcid.org/0000-0002-7008-6394)
- Isidora Glišić
- Špela Arhar Holdt (ORCID: https://orcid.org/0000-0003-0565-0531)
- Elena Volodina (ORCID: https://orcid.org/0000-0003-1935-1321)
- Ricardo Muñoz Sánchez (ORCID: https://orcid.org/0000-0002-9902-2925)
- Alexandros Tantos (ORCID: https://orcid.org/0000-0001-9917-1878)
- Ilze Auziņa (ORCID: https://orcid.org/0000-0001-6143-2841)
- Alexandr Rosen (ORCID: https://orcid.org/0000-0003-2210-4268)
- Alla Rozovskaya
- Andrew Caines (ORCID: https://orcid.org/0000-0001-9647-4902)
- Torsten Zesch (ORCID: https://orcid.org/0000-0002-9678-3825)
- Kais Allkivi-Metsoja (ORCID: https://orcid.org/0000-0003-3975-5104)
- Oleksiy Syvokon
- Orphée De Clercq (ORCID: https://orcid.org/0000-0002-6090-5552)
- Robert Östling (ORCID: https://orcid.org/0000-0002-5099-2584)
- Arianna Masciolini
- Mariana Romanyshyn
- Elena Drakonaki
- Pinelopi Kikilintza
- Despoina-Ourania Touriki
- Konstantinos Tsiotskas
- Eleni Tsourilla
- Vassilis Varsamopoulos
- Joni Kruijsbergen
Institutions
- FernUniversität in Hagen (DE)
- Institute of the Estonian Language (EE)
- Eurac Research (IT)
- University of Ljubljana (SI)
- Stockholm University (SE)
- University of Iceland (IS)
- Free University of Bozen-Bolzano (IT)
- City University of New York (US)
- Luleå University of Technology (SE)
- Tallinn University (EE)
- Charles University (CZ)
- Aristotle University of Thessaloniki (GR)
- University of Cambridge (GB)
- RISE Research Institutes of Sweden (SE)
- Ghent University Hospital (BE)
- Grammar School (SK)
- Microsoft Research (United Kingdom) (GB)
- University of Gothenburg (SE)
- University of Latvia (LV)
- Leipzig University (DE)
- University for Foreigners Perugia (IT)
Publication Details
- Journal
- International Journal of Learner Corpus Research
- Published
- 2025-04-01
- DOI
- https://doi.org/10.1075/ijlcr.24033.mas
- Citations
- 1
- Primary Topic
- Topic Modeling
- Type
- article
- Field-Weighted Citation Impact
- 2.48
Funders
- European Commission
- Vetenskapsrådet
- Latvijas Zinātnes Padome
- Latvijas Universitate
- Háskóli Íslands
- Aristotle University of Thessaloniki
- Univerza v Ljubljani
- Hellenic Foundation for Research and Innovation
- Research Committee, Aristotle University of Thessaloniki