Levenshtein Training for Word-level Quality Estimation

Shuoyang Ding, Marcin Junczys-Dowmunt, Matt Post, Philipp Koehn


Abstract
We propose a novel scheme to use the Levenshtein Transformer to perform the task of word-level quality estimation. A Levenshtein Transformer is a natural fit for this task: trained to perform decoding in an iterative manner, a Levenshtein Transformer can learn to post-edit without explicit supervision. To further minimize the mismatch between the translation task and the word-level QE task, we propose a two-stage transfer learning procedure on both augmented data and human post-editing data. We also propose heuristics to construct reference labels that are compatible with subword-level finetuning and inference. Results on WMT 2020 QE shared task dataset show that our proposed method has superior data efficiency under the data-constrained setting and competitive performance under the unconstrained setting.
Anthology ID:
2021.emnlp-main.539
Volume:
Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2021
Address:
Online and Punta Cana, Dominican Republic
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
6724–6733
Language:
URL:
https://aclanthology.org/2021.emnlp-main.539
DOI:
10.18653/v1/2021.emnlp-main.539
Bibkey:
Cite (ACL):
Shuoyang Ding, Marcin Junczys-Dowmunt, Matt Post, and Philipp Koehn. 2021. Levenshtein Training for Word-level Quality Estimation. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 6724–6733, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
Levenshtein Training for Word-level Quality Estimation (Ding et al., EMNLP 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.emnlp-main.539.pdf
Video:
 https://aclanthology.org/2021.emnlp-main.539.mp4
Code
 shuoyangd/stenella