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 machine translation (NAT) models have lower translationquality than autoregressive translation (AT) models because NAT decoders do notdepend on previous target tokens in the decoder input. We propose a novel andgeneral Dependency-Aware Decoder (DePA) to enhance target dependency modelingin the decoder of fully NAT models from two perspectives: decoderself-attention and decoder input. First, we propose an autoregressiveforward-backward pre-training phase before NAT training, which enables the NATdecoder to gradually learn bidirectional target dependencies for the final NATtraining. Second, we transform the decoder input from the source languagerepresentation space to the target language representation space through anovel attentive transformation process, which enables the decoder to bettercapture target dependencies. DePA can be applied to any fully NAT models.Extensive experiments show that DePA consistently improves highly competitiveand state-of-the-art fully NAT models on widely used WMT and IWSLT benchmarksby up to 1.88 BLEU gain, while maintaining the inference latency comparable toother fully NAT models.

DePA: Improving Non-autoregressive Machine Translation with Dependency-Aware Decoder