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

Existing techniques often attempt to make knowledge transfer from a powerfulmachine translation (MT) to speech translation (ST) model with some elaboratetechniques, which often requires transcription as extra input during training.However, transcriptions are not always available, and how to improve the STmodel performance without transcription, i.e., data efficiency, has rarely beenstudied in the literature. In this paper, we propose Decoupled Non-parametricKnowledge Distillation (DNKD) from data perspective to improve the dataefficiency. Our method follows the knowledge distillation paradigm. However,instead of obtaining the teacher distribution from a sophisticated MT model, weconstruct it from a non-parametric datastore via k-Nearest-Neighbor (kNN)retrieval, which removes the dependence on transcription and MT model. Then wedecouple the classic knowledge distillation loss into target and non-targetdistillation to enhance the effect of the knowledge among non-target logits,which is the prominent "dark knowledge". Experiments on MuST-C corpus showthat, the proposed method can achieve consistent improvement over the strongbaseline without requiring any transcription.

Decouple Non-parametric Knowledge Distillation For End-to-end Speech Translation