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

Speaker-independent speech recognition systems trained with data from manyusers are generally robust against speaker variability and work well for alarge population of speakers. However, these systems do not always generalizewell for users with very different speech characteristics. This issue can beaddressed by building personalized systems that are designed to work well foreach specific user. In this paper, we investigate the idea of securely trainingpersonalized end-to-end speech recognition models on mobile devices so thatuser data and models never leave the device and are never stored on a server.We study how the mobile training environment impacts performance by simulatingon-device data consumption. We conduct experiments using data collected fromspeech impaired users for personalization. Our results show thatpersonalization achieved 63.7\% relative word error rate reduction when trainedin a server environment and 58.1% in a mobile environment. Moving to on-devicepersonalization resulted in 18.7% performance degradation, in exchange forimproved scalability and data privacy. To train the model on device, we splitthe gradient computation into two and achieved 45% memory reduction at theexpense of 42% increase in training time.

An Investigation Into On-device Personalization of End-to-end Automatic Speech Recognition Models