Improving Fluency of Non-Autoregressive Machine Translation
This work addresses fluency issues in non-autoregressive machine translation, offering a more efficient solution for real-time translation applications, though it is incremental in nature.
The authors tackled the problem of impaired fluency in non-autoregressive machine translation models by enhancing beam search decoding with additional features, resulting in competitive BLEU scores while maintaining faster decoding speeds than autoregressive models across three language pairs.
Non-autoregressive (nAR) models for machine translation (MT) manifest superior decoding speed when compared to autoregressive (AR) models, at the expense of impaired fluency of their outputs. We improve the fluency of a nAR model with connectionist temporal classification (CTC) by employing additional features in the scoring model used during beam search decoding. Since the beam search decoding in our model only requires to run the network in a single forward pass, the decoding speed is still notably higher than in standard AR models. We train models for three language pairs: German, Czech, and Romanian from and into English. The results show that our proposed models can be more efficient in terms of decoding speed and still achieve a competitive BLEU score relative to AR models.