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

This paper describes a speaker diarization model based on target speakervoice activity detection (TS-VAD) using transformers. To overcome the originalTS-VAD model's drawback of being unable to handle an arbitrary number ofspeakers, we investigate model architectures that use input tensors withvariable-length time and speaker dimensions. Transformer layers are applied tothe speaker axis to make the model output insensitive to the order of thespeaker profiles provided to the TS-VAD model. Time-wise sequential layers areinterspersed between these speaker-wise transformer layers to allow thetemporal and cross-speaker correlations of the input speech signal to becaptured. We also extend a diarization model based on end-to-end neuraldiarization with encoder-decoder based attractors (EEND-EDA) by replacing itsdot-product-based speaker detection layer with the transformer-based TS-VAD.Experimental results on VoxConverse show that using the transformers for thecross-speaker modeling reduces the diarization error rate (DER) of TS-VAD by11.3%, achieving a new state-of-the-art (SOTA) DER of 4.57%. Also, our extendedEEND-EDA reduces DER by 6.9% on the CALLHOME dataset relative to the originalEEND-EDA with a similar model size, achieving a new SOTA DER of 11.18% under awidely used training data setting.

Target Speaker Voice Activity Detection with Transformers and Its Integration with End-to-End Neural Diarization