Using Optimal Transport as Alignment Objective for fine-tuning Multilingual Contextualized Embeddings

Sawsan Alqahtani, Garima Lalwani, Yi Zhang, Salvatore Romeo, Saab Mansour


Abstract
Recent studies have proposed different methods to improve multilingual word representations in contextualized settings including techniques that align between source and target embedding spaces. For contextualized embeddings, alignment becomes more complex as we additionally take context into consideration. In this work, we propose using Optimal Transport (OT) as an alignment objective during fine-tuning to further improve multilingual contextualized representations for downstream cross-lingual transfer. This approach does not require word-alignment pairs prior to fine-tuning that may lead to sub-optimal matching and instead learns the word alignments within context in an unsupervised manner. It also allows different types of mappings due to soft matching between source and target sentences. We benchmark our proposed method on two tasks (XNLI and XQuAD) and achieve improvements over baselines as well as competitive results compared to similar recent works.
Anthology ID:
2021.findings-emnlp.329
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2021
Month:
November
Year:
2021
Address:
Punta Cana, Dominican Republic
Editors:
Marie-Francine Moens, Xuanjing Huang, Lucia Specia, Scott Wen-tau Yih
Venue:
Findings
SIG:
SIGDAT
Publisher:
Association for Computational Linguistics
Note:
Pages:
3904–3919
Language:
URL:
https://aclanthology.org/2021.findings-emnlp.329
DOI:
10.18653/v1/2021.findings-emnlp.329
Bibkey:
Cite (ACL):
Sawsan Alqahtani, Garima Lalwani, Yi Zhang, Salvatore Romeo, and Saab Mansour. 2021. Using Optimal Transport as Alignment Objective for fine-tuning Multilingual Contextualized Embeddings. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 3904–3919, Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
Using Optimal Transport as Alignment Objective for fine-tuning Multilingual Contextualized Embeddings (Alqahtani et al., Findings 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.findings-emnlp.329.pdf
Video:
 https://aclanthology.org/2021.findings-emnlp.329.mp4
Data
XQuADXTREME