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 proposes an alternative algorithm for multichannel variationalautoencoder (MVAE), a recently proposed multichannel source separationapproach. While MVAE is notable in its impressive source separationperformance, the convergence-guaranteed optimization algorithm and that itallows us to estimate source-class labels simultaneously with sourceseparation, there are still two major drawbacks, i.e., the high computationalcomplexity and unsatisfactory source classification accuracy. To overcome thesedrawbacks, the proposed method employs an auxiliary classifier VAE, aninformation-theoretic extension of the conditional VAE, for learning thegenerative model of the source spectrograms. Furthermore, with the trainedauxiliary classifier, we introduce a novel algorithm for the optimization thatis able to not only reduce the computational time but also improve the sourceclassification performance. We call the proposed method "fast MVAE (fMVAE)".Experimental evaluations revealed that fMVAE achieved comparative sourceseparation performance to MVAE and about 80% source classification accuracyrate while it reduced about 93% computational time.

Fast MVAE: Joint separation and classification of mixed sources based on multichannel variational autoencoder with auxiliary classifier