Reinforcement Learning for Edit-Based Non-Autoregressive Neural Machine Translation

Hao Wang, Tetsuro Morimura, Ukyo Honda, Daisuke Kawahara


Abstract
Non-autoregressive (NAR) language models are known for their low latency in neural machine translation (NMT). However, a performance gap exists between NAR and autoregressive models due to the large decoding space and difficulty in capturing dependency between target words accurately. Compounding this, preparing appropriate training data for NAR models is a non-trivial task, often exacerbating exposure bias. To address these challenges, we apply reinforcement learning (RL) to Levenshtein Transformer, a representative edit-based NAR model, demonstrating that RL with self-generated data can enhance the performance of edit-based NAR models. We explore two RL approaches: stepwise reward maximization and episodic reward maximization. We discuss the respective pros and cons of these two approaches and empirically verify them. Moreover, we experimentally investigate the impact of temperature setting on performance, confirming the importance of proper temperature setting for NAR models’ training.
Anthology ID:
2024.naacl-srw.22
Volume:
Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 4: Student Research Workshop)
Month:
June
Year:
2024
Address:
Mexico City, Mexico
Editors:
Yang (Trista) Cao, Isabel Papadimitriou, Anaelia Ovalle
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
212–218
Language:
URL:
https://aclanthology.org/2024.naacl-srw.22
DOI:
Bibkey:
Cite (ACL):
Hao Wang, Tetsuro Morimura, Ukyo Honda, and Daisuke Kawahara. 2024. Reinforcement Learning for Edit-Based Non-Autoregressive Neural Machine Translation. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 4: Student Research Workshop), pages 212–218, Mexico City, Mexico. Association for Computational Linguistics.
Cite (Informal):
Reinforcement Learning for Edit-Based Non-Autoregressive Neural Machine Translation (Wang et al., NAACL 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.naacl-srw.22.pdf