English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Non-autoregressive approaches aim to improve the inference speed oftranslation models, particularly those that generate output in a one-passforward manner. However, these approaches often suffer from a significant dropin translation quality compared to autoregressive models. This paper introducesa series of innovative techniques to enhance the translation quality ofNon-Autoregressive Translation (NAT) models while maintaining a substantialacceleration in inference speed. We propose fine-tuning Pretrained MultilingualLanguage Models (PMLMs) with the CTC loss to train NAT models effectively.Furthermore, we adopt the MASK insertion scheme for up-sampling instead oftoken duplication, and we present an embedding distillation method to furtherenhance performance. In our experiments, our model outperforms the baselineautoregressive model (Transformer \textit{base}) on multiple datasets,including WMT'14 DE$\leftrightarrow$EN, WMT'16 RO$\leftrightarrow$EN, andIWSLT'14 DE$\leftrightarrow$EN. Notably, our model achieves better performancethan the baseline autoregressive model on the IWSLT'14 En$\leftrightarrow$Deand WMT'16 En$\leftrightarrow$Ro datasets, even without using distillation dataduring training. It is worth highlighting that on the IWSLT'14DE$\rightarrow$EN dataset, our model achieves an impressive BLEU score of39.59, setting a new state-of-the-art performance. Additionally, our modelexhibits a remarkable speed improvement of 16.35 times compared to theautoregressive model.

Improving Non-autoregressive Translation Quality with Pretrained Language Model, Embedding Distillation and Upsampling Strategy for CTC