On Efficiently Acquiring Annotations for Multilingual Models

Joel Moniz, Barun Patra, Matthew Gormley


Abstract
When tasked with supporting multiple languages for a given problem, two approaches have arisen: training a model for each language with the annotation budget divided equally among them, and training on a high-resource language followed by zero-shot transfer to the remaining languages. In this work, we show that the strategy of joint learning across multiple languages using a single model performs substantially better than the aforementioned alternatives. We also demonstrate that active learning provides additional, complementary benefits. We show that this simple approach enables the model to be data efficient by allowing it to arbitrate its annotation budget to query languages it is less certain on. We illustrate the effectiveness of our proposed method on a diverse set of tasks: a classification task with 4 languages, a sequence tagging task with 4 languages and a dependency parsing task with 5 languages. Our proposed method, whilst simple, substantially outperforms the other viable alternatives for building a model in a multilingual setting under constrained budgets.
Anthology ID:
2022.acl-short.9
Volume:
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
Month:
May
Year:
2022
Address:
Dublin, Ireland
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
69–85
Language:
URL:
https://aclanthology.org/2022.acl-short.9
DOI:
10.18653/v1/2022.acl-short.9
Bibkey:
Cite (ACL):
Joel Moniz, Barun Patra, and Matthew Gormley. 2022. On Efficiently Acquiring Annotations for Multilingual Models. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 69–85, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
On Efficiently Acquiring Annotations for Multilingual Models (Moniz et al., ACL 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.acl-short.9.pdf
Code
 codedecde/smal