AligNART: Non-autoregressive Neural Machine Translation by Jointly Learning to Estimate Alignment and Translate

Jongyoon Song, Sungwon Kim, Sungroh Yoon


Abstract
Non-autoregressive neural machine translation (NART) models suffer from the multi-modality problem which causes translation inconsistency such as token repetition. Most recent approaches have attempted to solve this problem by implicitly modeling dependencies between outputs. In this paper, we introduce AligNART, which leverages full alignment information to explicitly reduce the modality of the target distribution. AligNART divides the machine translation task into (i) alignment estimation and (ii) translation with aligned decoder inputs, guiding the decoder to focus on simplified one-to-one translation. To alleviate the alignment estimation problem, we further propose a novel alignment decomposition method. Our experiments show that AligNART outperforms previous non-iterative NART models that focus on explicit modality reduction on WMT14 En↔De and WMT16 Ro→En. Furthermore, AligNART achieves BLEU scores comparable to those of the state-of-the-art connectionist temporal classification based models on WMT14 En↔De. We also observe that AligNART effectively addresses the token repetition problem even without sequence-level knowledge distillation.
Anthology ID:
2021.emnlp-main.1
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:
1–14
Language:
URL:
https://aclanthology.org/2021.emnlp-main.1
DOI:
10.18653/v1/2021.emnlp-main.1
Bibkey:
Cite (ACL):
Jongyoon Song, Sungwon Kim, and Sungroh Yoon. 2021. AligNART: Non-autoregressive Neural Machine Translation by Jointly Learning to Estimate Alignment and Translate. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1–14, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
AligNART: Non-autoregressive Neural Machine Translation by Jointly Learning to Estimate Alignment and Translate (Song et al., EMNLP 2021)
Copy Citation:
PDF:
https://preview.aclanthology.org/emnlp-22-attachments/2021.emnlp-main.1.pdf
Video:
 https://preview.aclanthology.org/emnlp-22-attachments/2021.emnlp-main.1.mp4