Multilingual Pixel Representations for Translation and Effective Cross-lingual Transfer

Elizabeth Salesky, Neha Verma, Philipp Koehn, Matt Post


Abstract
We introduce and demonstrate how to effectively train multilingual machine translation models with pixel representations. We experiment with two different data settings with a variety of language and script coverage, demonstrating improved performance compared to subword embeddings. We explore various properties of pixel representations such as parameter sharing within and across scripts to better understand where they lead to positive transfer. We observe that these properties not only enable seamless cross-lingual transfer to unseen scripts, but make pixel representations more data-efficient than alternatives such as vocabulary expansion. We hope this work contributes to more extensible multilingual models for all languages and scripts.
Anthology ID:
2023.emnlp-main.854
Volume:
Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
13845–13861
Language:
URL:
https://aclanthology.org/2023.emnlp-main.854
DOI:
10.18653/v1/2023.emnlp-main.854
Bibkey:
Cite (ACL):
Elizabeth Salesky, Neha Verma, Philipp Koehn, and Matt Post. 2023. Multilingual Pixel Representations for Translation and Effective Cross-lingual Transfer. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 13845–13861, Singapore. Association for Computational Linguistics.
Cite (Informal):
Multilingual Pixel Representations for Translation and Effective Cross-lingual Transfer (Salesky et al., EMNLP 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.emnlp-main.854.pdf
Video:
 https://aclanthology.org/2023.emnlp-main.854.mp4