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/

Zendy Plus

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

Zendy Tools

Zendy Open

We use dynamic time warping (DTW) as supervision for training a convolutionalneural network (CNN) based keyword spotting system using a small set of spokenisolated keywords. The aim is to allow rapid deployment of a keyword spottingsystem in a new language to support urgent United Nations (UN) reliefprogrammes in parts of Africa where languages are extremely under-resourced andthe development of annotated speech resources is infeasible. First, we use 1920recorded keywords (40 keyword types, 34 minutes of speech) as exemplars in aDTW-based template matching system and apply it to untranscribed broadcastspeech. Then, we use the resulting DTW scores as targets to train a CNN on thesame unlabelled speech. In this way we use just 34 minutes of labelled speech,but leverage a large amount of unlabelled data for training. While theresulting CNN keyword spotter cannot match the performance of the DTW-basedsystem, it substantially outperforms a CNN classifier trained only on thekeywords, improving the area under the ROC curve from 0.54 to 0.64. Because ourCNN system is several orders of magnitude faster at runtime than the DTWsystem, it represents the most viable keyword spotter on this extremely limiteddataset.

Fast ASR-free and almost zero-resource keyword spotting using DTW and CNNs for humanitarian monitoring