The Volctrans GLAT System: Non-autoregressive Translation Meets WMT21
This provides a practical, fast parallel translation system for language processing tasks, though it appears incremental as it builds on existing non-autoregressive methods.
The paper tackles machine translation by building a non-autoregressive system using the Glancing Transformer for German->English translation in WMT21, achieving a BLEU score of 35.0 and outperforming all autoregressive models.
This paper describes the Volctrans' submission to the WMT21 news translation shared task for German->English translation. We build a parallel (i.e., non-autoregressive) translation system using the Glancing Transformer, which enables fast and accurate parallel decoding in contrast to the currently prevailing autoregressive models. To the best of our knowledge, this is the first parallel translation system that can be scaled to such a practical scenario like WMT competition. More importantly, our parallel translation system achieves the best BLEU score (35.0) on German->English translation task, outperforming all strong autoregressive counterparts.