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

Personal Digital Assistants (PDAs) - such as Siri, Alexa and GoogleAssistant, to name a few - play an increasingly important role to accessinformation and complete tasks spanning multiple domains, and by diverse groupsof users. A text-to-speech (TTS) module allows PDAs to interact in a natural,human-like manner, and play a vital role when the interaction involves peoplewith visual impairments or other disabilities. To cater to the needs of adiverse set of users, inclusive TTS is important to recognize and pronouncecorrectly text in different languages and dialects. Despite great progress inspeech synthesis, the pronunciation accuracy of named entities in amulti-lingual setting still has a large room for improvement. Existingapproaches to correct named entity (NE) mispronunciations, like retrainingGrapheme-to-Phoneme (G2P) models, or maintaining a TTS pronunciationdictionary, require expensive annotation of the ground truth pronunciation,which is also time consuming. In this work, we present a highly-precise,PDA-compatible pronunciation learning framework for the task of TTSmispronunciation detection and correction. In addition, we also propose a novelmispronunciation detection model called DTW-SiameseNet, which employs metriclearning with a Siamese architecture for Dynamic Time Warping (DTW) withtriplet loss. We demonstrate that a locale-agnostic, privacy-preservingsolution to the problem of TTS mispronunciation detection is feasible. Weevaluate our approach on a real-world dataset, and a corpus of NEpronunciations of an anonymized audio dataset of person names recorded byparticipants from 10 different locales. Human evaluation shows our proposedapproach improves pronunciation accuracy on average by ~6% compared to strongphoneme-based and audio-based baselines.

DTW-SiameseNet: Dynamic Time Warped Siamese Network for Mispronunciation Detection and Correction