Learning Metrics from Mean Teacher: A Supervised Learning Method for Improving the Generalization of Speaker Verification System

Abstract

Most speaker verification tasks are studied as an open-set evaluation scenario considering the real-world condition. Thus, the generalization power to unseen speakers is of paramount important to the performance of the speaker verification system. We propose to apply \textit {Mean Teacher}, a temporal averaging model, to extract speaker embeddings with small intra-class variance and large inter-class variance. The mean teacher network refers to the temporal averaging of deep neural network parameters; it can produces more accurate and stable representations than using weights after the training finished. By learning the reliable intermediate representation of the mean teacher network, we expect that the proposed method can explore more discriminatory embedding spaces and improve the generalization performance of the speaker verification system. Experimental results on the VoxCeleb1 test set demonstrate that the proposed method relatively improves performance by 11.61\%, compared to a baseline system.

5 pages, 1 figures, 5 tables, submitted to 2021 Interspeech as a conference paper

References in corpus (22)

• Adam: A Method for Stochastic Optimization
• Distilling the Knowledge in a Neural Network
• mixup: Beyond Empirical Risk Minimization
• SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
• Bootstrap your own latent: A new approach to self-supervised Learning
• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• VoxCeleb2: Deep Speaker Recognition
• VoxCeleb: a large-scale speaker identification dataset
• SGDR: Stochastic Gradient Descent with Warm Restarts
• On the Convergence of Adam and Beyond
• Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
• Attentive Statistics Pooling for Deep Speaker Embedding
• Large Batch Training of Convolutional Networks
• In defence of metric learning for speaker recognition
• CosFace: Large Margin Cosine Loss for Deep Face Recognition
• Ensemble Distillation for Neural Machine Translation
• Raw Waveform-based Speech Enhancement by Fully Convolutional Networks
• Meta-Learning for Short Utterance Speaker Recognition with Imbalance Length Pairs
• RawNet: Advanced end-to-end deep neural network using raw waveforms for text-independent speaker verification
• Speaker Recognition Based on Deep Learning: An Overview
• Improving Embedding Extraction for Speaker Verification with Ladder Network
• Open-set Short Utterance Forensic Speaker Verification using Teacher-Student Network with Explicit Inductive Bias