Cluster-Guided Unsupervised Domain Adaptation for Deep Speaker Embedding

Haiquan Mao, Feng Hong, Man wai Mak

Research output: Journal article publicationJournal articleAcademic researchpeer-review

1 Citation (Scopus)

Abstract

Recent studies have shown that pseudo labels can contribute to unsupervised domain adaptation (UDA) for speaker verification. Inspired by the self-training strategies that use an existing classifier to label the unlabeled data for retraining, we propose a cluster-guided UDA framework that labels the target domain data by clustering and combines the labeled source domain data and pseudo-labeled target domain data to train a speaker embedding network. To improve the cluster quality, we train a speaker embedding network dedicated for clustering by minimizing the contrastive center loss. The goal is to reduce the distance between an embedding and its assigned cluster center while enlarging the distance between the embedding and the other cluster centers. Using VoxCeleb2 as the source domain and CN-Celeb1 as the target domain, we demonstrate that the proposed method can achieve an equal error rate (EER) of 8.10% on the CN-Celeb1 evaluation set without using any labels from the target domain. This result outperforms the supervised baseline by 39.6% and is the state-of-the-art UDA performance on this corpus.

Original languageEnglish
Article number2303.15944
Pages (from-to)643-647
Number of pages5
JournalIEEE Signal Processing Letters
DOIs
Publication statusAccepted/In press - May 2023

Keywords

  • speaker verification
  • unsupervised domain adaptation
  • speaker clustering
  • Self-supervised learning
  • contrastive center loss

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Cluster-Guided Unsupervised Domain Adaptation for Deep Speaker Embedding'. Together they form a unique fingerprint.

Cite this