Abstract
We propose to train a non-autoregressive machine translation model to minimize the energy defined by a pretrained autoregressive model. In particular, we view our non-autoregressive translation system as an inference network (Tu and Gimpel, 2018) trained to minimize the autoregressive teacher energy. This contrasts with the popular approach of training a non-autoregressive model on a distilled corpus consisting of the beam-searched outputs of such a teacher model. Our approach, which we call ENGINE (ENerGy-based Inference NEtworks), achieves state-of-the-art non-autoregressive results on the IWSLT 2014 DE-EN and WMT 2016 RO-EN datasets, approaching the performance of autoregressive models.- Anthology ID:
- 2020.acl-main.251
- Volume:
- Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics
- Month:
- July
- Year:
- 2020
- Address:
- Online
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 2819–2826
- Language:
- URL:
- https://aclanthology.org/2020.acl-main.251
- DOI:
- 10.18653/v1/2020.acl-main.251
- Cite (ACL):
- Lifu Tu, Richard Yuanzhe Pang, Sam Wiseman, and Kevin Gimpel. 2020. ENGINE: Energy-Based Inference Networks for Non-Autoregressive Machine Translation. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 2819–2826, Online. Association for Computational Linguistics.
- Cite (Informal):
- ENGINE: Energy-Based Inference Networks for Non-Autoregressive Machine Translation (Tu et al., ACL 2020)
- PDF:
- https://preview.aclanthology.org/nodalida-main-page/2020.acl-main.251.pdf
- Code
- lifu-tu/ENGINE