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

Training a high performance end-to-end speech (E2E) processing model requiresan enormous amount of labeled speech data, especially in the era ofdata-centric artificial intelligence. However, labeled speech data are usuallyscarcer and more expensive for collection, compared to textual data. We proposeLatent Synthesis (LaSyn), an efficient textual data utilization framework forE2E speech processing models. We train a latent synthesizer to convert textualdata into an intermediate latent representation of a pre-trained speech model.These pseudo acoustic representations of textual data augment acoustic data formodel training. We evaluate LaSyn on low-resource automatic speech recognition(ASR) and spoken language understanding (SLU) tasks. For ASR, LaSyn improves anE2E baseline trained on LibriSpeech train-clean-100, with relative word errorrate reductions over 22.3% on different test sets. For SLU, LaSyn improves ourE2E baseline by absolute 4.1% for intent classification accuracy and 3.8% forslot filling SLU-F1 on SLURP, and absolute 4.49% and 2.25% for exact match (EM)and EM-Tree accuracies on STOP respectively. With fewer parameters, the resultsof LaSyn are competitive to published state-of-the-art works. The resultsdemonstrate the quality of the augmented training data.

Improving End-to-End Speech Processing by Efficient Text Data Utilization with Latent Synthesis