Additive Margin in Contrastive Self-Supervised Frameworks to Learn Discriminative Speaker Representations

arXiv:2404.14913v1 Announce Type: cross
Abstract: Self-Supervised Learning (SSL) frameworks became the standard for learning robust class representations by benefiting from large unlabeled datasets. For Speaker Verification (SV), most SSL systems rely on contrastive-based loss functions. We explore different ways to improve the performance of these techniques by revisiting the NT-Xent contrastive loss. Our main contribution is the definition of the NT-Xent-AM loss and the study of the importance of Additive Margin (AM) in SimCLR and MoCo SSL methods to further …

abstract arxiv class cs.lg cs.sd datasets eess.as explore frameworks functions learn loss performance robust self-supervised learning speaker ssl standard supervised learning systems type verification