Communication Dans Un Congrès Année : 2020

A Comparison of Metric Learning Loss Functions for End-To-End Speaker Verification

Résumé

Despite the growing popularity of metric learningapproaches, very little work has attempted to perform a fair comparison of these techniques for speaker verification. We try to fill this gap and compare several metric learning loss functions in a systematic manner on the VoxCeleb dataset. The first family of loss functions is derived from the cross entropy loss (usually used for supervised classification) and includes the congenerous cosine loss, the additive angular margin loss, and the center loss. The second family of loss functions focuses on the similarity between training samples and includes the contrastive loss and the triplet loss. We show that the additive angular margin loss function outperforms all other loss functions in the study, while learning more robust representations. Based on a combination of SincNet trainable features and the x-vector architecture, the network used in this paper brings us a step closer to a truly end-to-end speaker verification system, when combined with the additive angular margin loss, while still being competitive with the x-vector baseline. In the spirit of reproducible research, we also release open source Python code for reproducing our results, and share pretrained PyTorch models on torch.hub that can be used either directly or after fine-tuning.

Dates et versions

hal-02989334 , version 1 (05-11-2020)

Identifiants

Citer

Juan Manuel Coria, Hervé Bredin, Sahar Ghannay, Sophie Rosset. A Comparison of Metric Learning Loss Functions for End-To-End Speaker Verification. International Conference on Statistical Language and Speech Processing, Oct 2020, Cardiff, United Kingdom. pp.137-148, ⟨10.1007/978-3-030-59430-5_11⟩. ⟨hal-02989334⟩
193 Consultations
0 Téléchargements

Altmetric

Partager

  • More