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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Learning meaningful and general representations from unannotated speech thatare applicable to a wide range of tasks remains challenging. In this paper wepropose to use autoregressive predictive coding (APC), a recently proposedself-supervised objective, as a generative pre-training approach for learningmeaningful, non-specific, and transferable speech representations. We pre-trainAPC on large-scale unlabeled data and conduct transfer learning experiments onthree speech applications that require different information about speechcharacteristics to perform well: speech recognition, speech translation, andspeaker identification. Extensive experiments show that APC not onlyoutperforms surface features (e.g., log Mel spectrograms) and other popularrepresentation learning methods on all three tasks, but is also effective atreducing downstream labeled data size and model parameters. We also investigatethe use of Transformers for modeling APC and find it superior to RNNs.

Generative Pre-Training for Speech with Autoregressive Predictive Coding