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/

Zendy Plus

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

Zendy Tools

Zendy Open

Conditional masked language models (CMLM) have shown impressive progress innon-autoregressive machine translation (NAT). They learn the conditionaltranslation model by predicting the random masked subset in the targetsentence. Based on the CMLM framework, we introduce Multi-view SubsetRegularization (MvSR), a novel regularization method to improve the performanceof the NAT model. Specifically, MvSR consists of two parts: (1) \textit{sharedmask consistency}: we forward the same target with different mask strategies,and encourage the predictions of shared mask positions to be consistent witheach other. (2) \textit{model consistency}, we maintain an exponential movingaverage of the model weights, and enforce the predictions to be consistentbetween the average model and the online model. Without changing the CMLM-basedarchitecture, our approach achieves remarkable performance on three publicbenchmarks with 0.36-1.14 BLEU gains over previous NAT models. Moreover,compared with the stronger Transformer baseline, we reduce the gap to 0.01-0.44BLEU scores on small datasets (WMT16 RO$\leftrightarrow$EN and IWSLTDE$\rightarrow$EN).

MvSR-NAT: Multi-view Subset Regularization for Non-Autoregressive Machine Translation