Label-efficient self-supervised speaker verification with information maximization and contrastive learning

Abstract