Language Models Are Polyglots: Language Similarity Predicts Cross-Lingual Transfer Learning Performance
Machine Learning and Knowledge Extraction · 07.03.2026 · 8 (3) · 65
DOI: 10.3390/make8030065 ↗Kurz gefasst
Fachartikel in der Zeitschrift Machine Learning and Knowledge Extraction, online erschienen am 07.03.2026. Verfasst von Juuso Eronen und 7 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 122 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Selecting a source language for zero-shot cross-lingual transfer is typically done by intuition or by defaulting to English, despite large performance differences across language pairs. We study whether linguistic similarity can predict transfer performance and support principled source-language selection. We introduce quantified WALS (qWALS), a typology-based similarity metric derived from features in the World Atlas of Language Structures, and evaluate it against existing similarity baselines. Validation uses three complementary signals: computational similarity scores, zero-shot transfer performance of multilingual transformers (mBERT and XLM-R) on four NLP tasks (dependency parsing, named entity recognition, sentiment analysis, and abusive language identification) across eight languages, and an expert-linguist similarity survey. Across tasks and models, higher linguistic similarity is associated with better transfer, and the survey provides independent support for the computational metrics.
Zitieren
APA
Eronen, J., Ptaszynski, M., Wicherkiewicz, T., Borges, R., Janic, K., Liu, Z., Mahmud, T., & Masui, F. (2026). Language Models Are Polyglots: Language Similarity Predicts Cross-Lingual Transfer Learning Performance. Machine Learning and Knowledge Extraction, 8(3), 65. https://doi.org/10.3390/make8030065
Vancouver
Eronen J, Ptaszynski M, Wicherkiewicz T, Borges R, Janic K, Liu Z, et al. Language Models Are Polyglots: Language Similarity Predicts Cross-Lingual Transfer Learning Performance. Machine Learning and Knowledge Extraction. 2026;8(3):65. doi:10.3390/make8030065
BibTeX
@article{Eronen2026,
author = {Eronen, Juuso and Ptaszynski, Michal and Wicherkiewicz, Tomasz and Borges, Robert and Janic, Katarzyna and Liu, Zhenzhen and Mahmud, Tanjim and Masui, Fumito},
title = {Language Models Are Polyglots: Language Similarity Predicts Cross-Lingual Transfer Learning Performance},
journal = {Machine Learning and Knowledge Extraction},
year = {2026},
volume = {8},
number = {3},
pages = {65},
doi = {10.3390/make8030065}
}
EUVIMED