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

Neural network-based speaker recognition has achieved significant improvementin recent years. A robust speaker representation learns meaningful knowledgefrom both hard and easy samples in the training set to achieve goodperformance. However, noisy samples (i.e., with wrong labels) in the trainingset induce confusion and cause the network to learn the incorrectrepresentation. In this paper, we propose a two-step audio-visual deepcleansing framework to eliminate the effect of noisy labels in speakerrepresentation learning. This framework contains a coarse-grained cleansingstep to search for the peculiar samples, followed by a fine-grained cleansingstep to filter out the noisy labels. Our study starts from an efficientaudio-visual speaker recognition system, which achieves a close to perfectequal-error-rate (EER) of 0.01\%, 0.07\% and 0.13\% on the Vox-O, E and H testsets. With the proposed multi-modal cleansing mechanism, four different speakerrecognition networks achieve an average improvement of 5.9\%. Code has beenmade available at:\textcolor{magenta}{\url{https://github.com/TaoRuijie/AVCleanse}}.

Speaker recognition with two-step multi-modal deep cleansing