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 (SSL) has shown significant progress in speechprocessing tasks. However, despite the intrinsic randomness in the Transformerstructure, such as dropout variants and layer-drop, improving the model-levelconsistency remains under-explored in the speech SSL literature. To addressthis, we propose a new pre-training method that uses consistency regularizationto improve Data2vec 2.0, the recent state-of-the-art (SOTA) SSL model.Specifically, the proposed method involves sampling two different studentsub-models within the Data2vec 2.0 framework, enabling two output variantsderived from a single input without additional parameters. Subsequently, weregularize the outputs from the student sub-models to be consistent and requirethem to predict the representation of the teacher model. Our experimentalresults demonstrate that the proposed approach improves the SSL model'srobustness and generalization ability, resulting in SOTA results on the SUPERBbenchmark.

MCR-Data2vec 2.0: Improving Self-supervised Speech Pre-training via Model-level Consistency Regularization