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

Self-supervised learning in speech involves training a speech representationnetwork on a large-scale unannotated speech corpus, and then applying thelearned representations to downstream tasks. Since the majority of thedownstream tasks of SSL learning in speech largely focus on the contentinformation in speech, the most desirable speech representations should be ableto disentangle unwanted variations, such as speaker variations, from thecontent. However, disentangling speakers is very challenging, because removingthe speaker information could easily result in a loss of content as well, andthe damage of the latter usually far outweighs the benefit of the former. Inthis paper, we propose a new SSL method that can achieve speakerdisentanglement without severe loss of content. Our approach is adapted fromthe HuBERT framework, and incorporates disentangling mechanisms to regularizeboth the teacher labels and the learned representations. We evaluate thebenefit of speaker disentanglement on a set of content-related downstreamtasks, and observe a consistent and notable performance advantage of ourspeaker-disentangled representations.

ContentVec: An Improved Self-Supervised Speech Representation by Disentangling Speakers