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

Many speech processing tasks involve measuring the acoustic similaritybetween speech segments. Acoustic word embeddings (AWE) allow for efficientcomparisons by mapping speech segments of arbitrary duration tofixed-dimensional vectors. For zero-resource speech processing, whereunlabelled speech is the only available resource, some of the best AWEapproaches rely on weak top-down constraints in the form of automaticallydiscovered word-like segments. Rather than learning embeddings at the segmentlevel, another line of zero-resource research has looked at representationlearning at the short-time frame level. Recent approaches includeself-supervised predictive coding and correspondence autoencoder (CAE) models.In this paper we consider whether these frame-level features are beneficialwhen used as inputs for training to an unsupervised AWE model. We compareframe-level features from contrastive predictive coding (CPC), autoregressivepredictive coding and a CAE to conventional MFCCs. These are used as inputs toa recurrent CAE-based AWE model. In a word discrimination task on English andXitsonga data, all three representation learning approaches outperform MFCCs,with CPC consistently showing the biggest improvement. In cross-lingualexperiments we find that CPC features trained on English can also betransferred to Xitsonga.

A comparison of self-supervised speech representations as input features for unsupervised acoustic word embeddings