Saved in:
Bibliographic Details
Main Authors: Chen, Yafeng, Deng, Chong, Wang, Hui, Jiang, Yiheng, Yin, Han, Chen, Qian, Wang, Wen
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.13826
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908371374833664
author Chen, Yafeng
Deng, Chong
Wang, Hui
Jiang, Yiheng
Yin, Han
Chen, Qian
Wang, Wen
author_facet Chen, Yafeng
Deng, Chong
Wang, Hui
Jiang, Yiheng
Yin, Han
Chen, Qian
Wang, Wen
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2505_13826
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Pushing the Frontiers of Self-Distillation Prototypes Network with Dimension Regularization and Score Normalization
Chen, Yafeng
Deng, Chong
Wang, Hui
Jiang, Yiheng
Yin, Han
Chen, Qian
Wang, Wen
Audio and Speech Processing
Sound
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.
title Pushing the Frontiers of Self-Distillation Prototypes Network with Dimension Regularization and Score Normalization
topic Audio and Speech Processing
Sound
url https://arxiv.org/abs/2505.13826