Accurate Online Posterior Alignments for Principled Lexically-Constrained Decoding

Soumya Chatterjee, Sunita Sarawagi, Preethi Jyothi


Abstract
Online alignment in machine translation refers to the task of aligning a target word to a source word when the target sequence has only been partially decoded. Good online alignments facilitate important applications such as lexically constrained translation where user-defined dictionaries are used to inject lexical constraints into the translation model. We propose a novel posterior alignment technique that is truly online in its execution and superior in terms of alignment error rates compared to existing methods. Our proposed inference technique jointly considers alignment and token probabilities in a principled manner and can be seamlessly integrated within existing constrained beam-search decoding algorithms. On five language pairs, including two distant language pairs, we achieve consistent drop in alignment error rates. When deployed on seven lexically constrained translation tasks, we achieve significant improvements in BLEU specifically around the constrained positions.
Anthology ID:
2022.acl-long.460
Volume:
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:
May
Year:
2022
Address:
Dublin, Ireland
Editors:
Smaranda Muresan, Preslav Nakov, Aline Villavicencio
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
6675–6689
Language:
URL:
https://aclanthology.org/2022.acl-long.460
DOI:
10.18653/v1/2022.acl-long.460
Bibkey:
Cite (ACL):
Soumya Chatterjee, Sunita Sarawagi, and Preethi Jyothi. 2022. Accurate Online Posterior Alignments for Principled Lexically-Constrained Decoding. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6675–6689, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
Accurate Online Posterior Alignments for Principled Lexically-Constrained Decoding (Chatterjee et al., ACL 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.acl-long.460.pdf
Video:
 https://aclanthology.org/2022.acl-long.460.mp4