Unsupervised Speaker Verification Using Pre-Trained Model and Label Correction

Zhicong Chen; Jie Wang; Wenxuan Hu; Lin Li; Qingyang Hong

ICASSP 2023accepted0 citations

Unsupervised Speaker Verification Using Pre-Trained Model and Label Correction

Zhicong Chen, Jie Wang, Wenxuan Hu, Lin Li, Qingyang Hong

Abstract

Recently, the fine-tuning pre-trained model framework has emerged as a promising paradigm for speech-processing tasks. In this study, we present a novel strategy for unsupervised speaker verification using the Sub-structure of Pre-Trained Model (Sub-PTM), which consists of a CNN-based feature extractor and several Transformer blocks. To obtain the initial pseudo labels, we utilize Infomap to perform clustering on the representations extracted from the Sub-PTM. The generated pseudo labels are then leveraged to train a speaker verification model containing a Sub-PTM and a downstream network. We also propose an Online and Offline Label Correction (OAO-LC) method to alleviate the effects of incorrect pseudo labels. By incorporating these techniques, our system achieves competitive results compared to the supervised baseline.

BibTeX

@inproceedings{icassp2023_unsupervisedspea,
  title = {Unsupervised Speaker Verification Using Pre-Trained Model and Label Correction},
  author = {Zhicong Chen and Jie Wang and Wenxuan Hu and Lin Li and Qingyang Hong},
  booktitle = {ICASSP 2023},
  year = {2023}
}