Guardado en:
Detalles Bibliográficos
Autores principales: Lepage, Theo, Dehak, Reda
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2505.14561
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909742488616960
author Lepage, Theo
Dehak, Reda
author_facet Lepage, Theo
Dehak, Reda
contents Self-Supervised Learning (SSL) has led to considerable progress in Speaker Verification (SV). The standard framework uses same-utterance positive sampling and data-augmentation to generate anchor-positive pairs of the same speaker. This is a major limitation, as this strategy primarily encodes channel information from the recording condition, shared by the anchor and positive. We propose a new positive sampling technique to address this bottleneck: Self-Supervised Positive Sampling (SSPS). For a given anchor, SSPS aims to find an appropriate positive, i.e., of the same speaker identity but a different recording condition, in the latent space using clustering assignments and a memory queue of positive embeddings. SSPS improves SV performance for both SimCLR and DINO, reaching 2.57% and 2.53% EER, outperforming SOTA SSL methods on VoxCeleb1-O. In particular, SimCLR-SSPS achieves a 58% EER reduction by lowering intra-speaker variance, providing comparable performance to DINO-SSPS.
format Preprint
id arxiv_https___arxiv_org_abs_2505_14561
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SSPS: Self-Supervised Positive Sampling for Robust Self-Supervised Speaker Verification
Lepage, Theo
Dehak, Reda
Audio and Speech Processing
Artificial Intelligence
Machine Learning
Sound
Self-Supervised Learning (SSL) has led to considerable progress in Speaker Verification (SV). The standard framework uses same-utterance positive sampling and data-augmentation to generate anchor-positive pairs of the same speaker. This is a major limitation, as this strategy primarily encodes channel information from the recording condition, shared by the anchor and positive. We propose a new positive sampling technique to address this bottleneck: Self-Supervised Positive Sampling (SSPS). For a given anchor, SSPS aims to find an appropriate positive, i.e., of the same speaker identity but a different recording condition, in the latent space using clustering assignments and a memory queue of positive embeddings. SSPS improves SV performance for both SimCLR and DINO, reaching 2.57% and 2.53% EER, outperforming SOTA SSL methods on VoxCeleb1-O. In particular, SimCLR-SSPS achieves a 58% EER reduction by lowering intra-speaker variance, providing comparable performance to DINO-SSPS.
title SSPS: Self-Supervised Positive Sampling for Robust Self-Supervised Speaker Verification
topic Audio and Speech Processing
Artificial Intelligence
Machine Learning
Sound
url https://arxiv.org/abs/2505.14561