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

The success of deep learning comes from its ability to capture thehierarchical structure of data by learning high-level representations definedin terms of low-level ones. In this paper we explore self-supervised learningof hierarchical representations of speech by applying multiple levels ofContrastive Predictive Coding (CPC). We observe that simply stacking two CPCmodels does not yield significant improvements over single-level architectures.Inspired by the fact that speech is often described as a sequence of discreteunits unevenly distributed in time, we propose a model in which the output of alow-level CPC module is non-uniformly downsampled to directly minimize the lossof a high-level CPC module. The latter is designed to also enforce a prior ofseparability and discreteness in its representations by enforcing dissimilarityof successive high-level representations through focused negative sampling, andby quantization of the prediction targets. Accounting for the structure of thespeech signal improves upon single-level CPC features and enhances thedisentanglement of the learned representations, as measured by downstreamspeech recognition tasks, while resulting in a meaningful segmentation of thesignal that closely resembles phone boundaries.

Variable-rate hierarchical CPC leads to acoustic unit discovery in speech