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

Speech fluency/disfluency can be evaluated by analyzing a range of phoneticand prosodic features. Deep neural networks are commonly trained to mapfluency-related features into the human scores. However, the effectiveness ofdeep learning-based models is constrained by the limited amount of labeledtraining samples. To address this, we introduce a self-supervised learning(SSL) approach that takes into account phonetic and prosody awareness forfluency scoring. Specifically, we first pre-train the model using areconstruction loss function, by masking phones and their durations jointly ona large amount of unlabeled speech and text prompts. We then fine-tune thepre-trained model using human-annotated scoring data. Our experimental results,conducted on datasets such as Speechocean762 and our non-native datasets, showthat our proposed method outperforms the baseline systems in terms of Pearsoncorrelation coefficients (PCC). Moreover, we also conduct an ablation study tobetter understand the contribution of phonetic and prosody factors during thepre-training stage.

Phonetic and Prosody-aware Self-supervised Learning Approach for Non-native Fluency Scoring