Too many cooks spoil the model : are bilingual models for Slovene better than a large multilingual model?

Publication type
C1
Publication status
Published
Authors
Singh, P., Maladry, A, & Lefever, E.
Editor
Jakub Piskorski, Michał Marcińczuk, Preslav Nakov, Maciej Ogrodniczuk, Senja Pollak, Pavel Přibáň, Piotr Rybak, Josef Steinberger and Roman Yangarber
Series
Proceedings of the 9th Workshop on Slavic Natural Language Processing 2023
Pagination
32-39
Publisher
Association for Computational Linguistics
Conference
Slav-NLP 2023 : The 9th Workshop on NLP for Slavic languages (Dubrovnik, Croatia)
Download
(.pdf)
View in Biblio
(externe link)

Abstract

This paper investigates whether adding data of typologically closer languages improves the performance of transformer-based models for three different downstream tasks, namely Part-of-Speech tagging, Named Entity Recognition, and Sentiment Analysis, compared to a monolingual and plain multilingual language model. For the presented pilot study, we performed experiments for the use case of Slovene, a low(er)-resourced language belonging to the Slavic language family. The experiments were carried out in a controlled setting, where a monolingual model for Slovene was compared to combined language models containing Slovene, trained with the same amount of Slovene data. The experimental results show that adding typologically closer languages indeed improves the performance of the Slovene language model, and even succeeds in outperforming the large multilingual XLM-RoBERTa model for NER and PoS-tagging. We also reveal that, contrary to intuition, distantly or unrelated languages also combine admirably with Slovene, often out-performing XLM-R as well. All the bilingual models used in the experiments are publicly available at https://github.com/pranaydeeps/BLAIR