Contrastive Speaker Embedding With Sequential Disentanglement

Contrastive speaker embedding assumes that the contrast between the positive and negative pairs of speech segments is attributed to speaker identity only. However, this assumption is incorrect because speech signals contain not only speaker identity but also linguistic content. In this paper, we pro...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings of the ... IEEE International Conference on Acoustics, Speech and Signal Processing (1998) pp. 10891 - 10895
Main Authors: Tu, Youzhi, Mak, Man-Wai, Chien, Jen-Tzung
Format: Conference Proceeding
Language:English
Published: IEEE 14.04.2024
Subjects:
Acoustics
disentangled representation learning
Linguistics
Self-supervised learning
Signal processing
SimCLR
speaker embedding
Speaker verification
Speech processing
VAE
ISSN:2379-190X
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Contrastive speaker embedding assumes that the contrast between the positive and negative pairs of speech segments is attributed to speaker identity only. However, this assumption is incorrect because speech signals contain not only speaker identity but also linguistic content. In this paper, we propose a contrastive learning framework with sequential disentanglement to remove linguistic content by incorporating a disentangled sequential variational autoencoder (DSVAE) into the conventional SimCLR framework. The DSVAE aims to disentangle speaker factors from content factors in an embedding space so that only the speaker factors are used for constructing a contrastive loss objective. Because content factors have been removed from the contrastive learning, the resulting speaker embeddings will be content-invariant. Experimental results on VoxCeleb1-test show that the proposed method consistently outperforms SimCLR. This suggests that applying sequential disentanglement is beneficial to learning speaker-discriminative embeddings.
ISSN:2379-190X
DOI:10.1109/ICASSP48485.2024.10448284