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 presents a description of STC Ltd. systems submitted to the NIST2021 Speaker Recognition Evaluation for both fixed and open trainingconditions. These systems consists of a number of diverse subsystems based onusing deep neural networks as feature extractors. During the NIST 2021 SREchallenge we focused on the training of the state-of-the-art deep speakerembeddings extractors like ResNets and ECAPA networks by using additive angularmargin based loss functions. Additionally, inspired by the recent success ofthe wav2vec 2.0 features in automatic speech recognition we explored theeffectiveness of this approach for the speaker verification filed. According toour observation the fine-tuning of the pretrained large wav2vec 2.0 modelprovides our best performing systems for open track condition. Our experimentswith wav2vec 2.0 based extractors for the fixed condition showed thatunsupervised autoregressive pretraining with Contrastive Predictive Coding lossopens the door to training powerful transformer-based extractors from rawspeech signals. For video modality we developed our best solution withRetinaFace face detector and deep ResNet face embeddings extractor trained onlarge face image datasets. The final results for primary systems were obtainedby different configurations of subsystems fusion on the score level followed byscore calibration.

STC speaker recognition systems for the NIST SRE 2021