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.
@article{arxiv.2005.00850,
title = {ENGINE: Energy-Based Inference Networks for Non-Autoregressive Machine Translation},
author = {Lifu Tu and Richard Yuanzhe Pang and Sam Wiseman and Kevin Gimpel},
journal= {arXiv preprint arXiv:2005.00850},
year = {2020}
}