Speaker-IPL: Unsupervised Learning of Speaker Characteristics with i-Vector based Pseudo-Labels

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Aldeneh, Zakaria, Higuchi, Takuya, Jung, Jee-weon, Chen, Li-Wei, Shum, Stephen, Abdelaziz, Ahmed Hussen, Watanabe, Shinji, Likhomanenko, Tatiana, Theobald, Barry-John
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910788866801664
author Aldeneh, Zakaria
Higuchi, Takuya
Jung, Jee-weon
Chen, Li-Wei
Shum, Stephen
Abdelaziz, Ahmed Hussen
Watanabe, Shinji
Likhomanenko, Tatiana
Theobald, Barry-John
author_facet Aldeneh, Zakaria
Higuchi, Takuya
Jung, Jee-weon
Chen, Li-Wei
Shum, Stephen
Abdelaziz, Ahmed Hussen
Watanabe, Shinji
Likhomanenko, Tatiana
Theobald, Barry-John
contents Iterative self-training, or iterative pseudo-labeling (IPL) -- using an improved model from the current iteration to provide pseudo-labels for the next iteration -- has proven to be a powerful approach to enhance the quality of speaker representations. Recent applications of IPL in unsupervised speaker recognition start with representations extracted from very elaborate self-supervised methods (e.g., DINO). However, training such strong self-supervised models is not straightforward (they require hyper-parameter tuning and may not generalize to out-of-domain data) and, moreover, may not be needed at all. To this end, we show that the simple, well-studied, and established i-vector generative model is enough to bootstrap the IPL process for the unsupervised learning of speaker representations. We also systematically study the impact of other components on the IPL process, which includes the initial model, the encoder, augmentations, the number of clusters, and the clustering algorithm. Remarkably, we find that even with a simple and significantly weaker initial model like i-vector, IPL can still achieve speaker verification performance that rivals state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2409_10791
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Speaker-IPL: Unsupervised Learning of Speaker Characteristics with i-Vector based Pseudo-Labels
Aldeneh, Zakaria
Higuchi, Takuya
Jung, Jee-weon
Chen, Li-Wei
Shum, Stephen
Abdelaziz, Ahmed Hussen
Watanabe, Shinji
Likhomanenko, Tatiana
Theobald, Barry-John
Audio and Speech Processing
Sound
Iterative self-training, or iterative pseudo-labeling (IPL) -- using an improved model from the current iteration to provide pseudo-labels for the next iteration -- has proven to be a powerful approach to enhance the quality of speaker representations. Recent applications of IPL in unsupervised speaker recognition start with representations extracted from very elaborate self-supervised methods (e.g., DINO). However, training such strong self-supervised models is not straightforward (they require hyper-parameter tuning and may not generalize to out-of-domain data) and, moreover, may not be needed at all. To this end, we show that the simple, well-studied, and established i-vector generative model is enough to bootstrap the IPL process for the unsupervised learning of speaker representations. We also systematically study the impact of other components on the IPL process, which includes the initial model, the encoder, augmentations, the number of clusters, and the clustering algorithm. Remarkably, we find that even with a simple and significantly weaker initial model like i-vector, IPL can still achieve speaker verification performance that rivals state-of-the-art methods.
title Speaker-IPL: Unsupervised Learning of Speaker Characteristics with i-Vector based Pseudo-Labels
topic Audio and Speech Processing
Sound
url https://arxiv.org/abs/2409.10791