Pushing the Frontiers of Self-Distillation Prototypes Network with Dimension Regularization and Score Normalization

TL;DR

This paper improves self-supervised speaker verification by introducing dimension regularization and score normalization to the SDPN framework, significantly narrowing the performance gap with supervised methods.

Contribution

It proposes a novel combination of dimension regularization and score normalization within SDPN, achieving state-of-the-art results in self-supervised speaker verification.

Findings

Achieved state-of-the-art EER on VoxCeleb1 benchmark

Improved self-supervised SV performance by over 20%

Effectively addressed embedding collapse with regularization

Abstract

Developing robust speaker verification (SV) systems without speaker labels has been a longstanding challenge. Earlier research has highlighted a considerable performance gap between self-supervised and fully supervised approaches. In this paper, we enhance the non-contrastive self-supervised framework, Self-Distillation Prototypes Network (SDPN), by introducing dimension regularization that explicitly addresses the collapse problem through the application of regularization terms to speaker embeddings. Moreover, we integrate score normalization techniques from fully supervised SV to further bridge the gap toward supervised verification performance. SDPN with dimension regularization and score normalization sets a new state-of-the-art on the VoxCeleb1 speaker verification evaluation benchmark, achieving Equal Error Rate 1.29%, 1.60%, and 2.80% for trial VoxCeleb1-{O,E,H} respectively. These results demonstrate relative improvements of 28.3%, 19.6%, and 22.6% over the current best self-supervised methods, thereby advancing the frontiers of SV technology.