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

In this paper, we address the problem of speaker recognition in challengingacoustic conditions using a novel method to extract robustspeaker-discriminative speech representations. We adopt a recently proposedunsupervised adversarial invariance architecture to train a network that mapsspeaker embeddings extracted using a pre-trained model onto two lowerdimensional embedding spaces. The embedding spaces are learnt to disentanglespeaker-discriminative information from all other information present in theaudio recordings, without supervision about the acoustic conditions. We analyzethe robustness of the proposed embeddings to various sources of variabilitypresent in the signal for speaker verification and unsupervised clusteringtasks on a large-scale speaker recognition corpus. Our analyses show that theproposed system substantially outperforms the baseline in a variety ofchallenging acoustic scenarios. Furthermore, for the task of speakerdiarization on a real-world meeting corpus, our system shows a relativeimprovement of 36\% in the diarization error rate compared to thestate-of-the-art baseline.

Robust speaker recognition using unsupervised adversarial invariance