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

Spoken language understanding (SLU) requires a model to analyze inputacoustic signal to understand its linguistic content and make predictions. Toboost the models' performance, various pre-training methods have been proposedto learn rich representations from large-scale unannotated speech and text.However, the inherent disparities between the two modalities necessitate amutual analysis. In this paper, we propose a novel semi-supervised learningframework, SPLAT, to jointly pre-train the speech and language modules. Besidesconducting a self-supervised masked language modeling task on the twoindividual modules using unpaired speech and text, SPLAT aligns representationsfrom the two modules in a shared latent space using a small amount of pairedspeech and text. Thus, during fine-tuning, the speech module alone can producerepresentations carrying both acoustic information and contextual semanticknowledge of an input acoustic signal. Experimental results verify theeffectiveness of our approach on various SLU tasks. For example, SPLAT improvesthe previous state-of-the-art performance on the Spoken SQuAD dataset by morethan 10%.

SPLAT: Speech-Language Joint Pre-Training for Spoken Language Understanding