Non-Autoregressive Document-Level Machine Translation
This addresses the challenge of using faster NAT models in real-world document translation scenarios, though it is incremental as it explores existing methods in a new context.
The paper tackled the problem of applying non-autoregressive translation (NAT) models to document-level machine translation, finding that NAT models achieve high acceleration but still have a significant performance gap compared to autoregressive models, with sentence alignment enhancing their performance.
Non-autoregressive translation (NAT) models achieve comparable performance and superior speed compared to auto-regressive translation (AT) models in the context of sentence-level machine translation (MT). However, their abilities are unexplored in document-level MT, hindering their usage in real scenarios. In this paper, we conduct a comprehensive examination of typical NAT models in the context of document-level MT and further propose a simple but effective design of sentence alignment between source and target. Experiments show that NAT models achieve high acceleration on documents, and sentence alignment significantly enhances their performance. However, current NAT models still have a significant performance gap compared to their AT counterparts. Further investigation reveals that NAT models suffer more from the multi-modality and misalignment issues in the context of document-level MT, and current NAT models struggle with exploiting document context and handling discourse phenomena. We delve into these challenges and provide our code at \url{https://github.com/baoguangsheng/nat-on-doc}.