Cluster-Guided Unsupervised Domain Adaptation for Deep Speaker Embedding

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.

References in corpus (10)

• Adam: A Method for Stochastic Optimization
• A Simple Framework for Contrastive Learning of Visual Representations
• ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification
• Additive Margin Softmax for Face Verification
• Deep Clustering for Unsupervised Learning of Visual Features
• In defence of metric learning for speaker recognition
• Clustering with Deep Learning: Taxonomy and New Methods
• Self-Supervised Learning Based Domain Adaptation for Robust Speaker Verification
• Attention Back-end for Automatic Speaker Verification with Multiple Enrollment Utterances
• TMS: A Temporal Multi-scale Backbone Design for Speaker Embedding