Channel adversarial training for speaker verification and diarization

Abstract

Previous work has encouraged domain-invariance in deep speaker embedding by adversarially classifying the dataset or labelled environment to which the generated features belong. We propose a training strategy which aims to produce features that are invariant at the granularity of the recording or channel, a finer grained objective than dataset- or environment-invariance. By training an adversary to predict whether pairs of same-speaker embeddings belong to the same recording in a Siamese fashion, learned features are discouraged from utilizing channel information that may be speaker discriminative during training. Experiments for verification on VoxCeleb and diarization and verification on CALLHOME show promising improvements over a strong baseline in addition to outperforming a dataset-adversarial model. The VoxCeleb model in particular performs well, achieving a $4\%$ relative improvement in EER over a Kaldi baseline, while using a similar architecture and less training data.

Submitted to IEEE ICASSP 2020

References in corpus (8)

• VoxCeleb2: Deep Speaker Recognition
• VoxCeleb: a large-scale speaker identification dataset
• Additive Margin Softmax for Face Verification
• Attentive Statistics Pooling for Deep Speaker Embedding
• Wasserstein Distance Guided Representation Learning for Domain Adaptation
• LSTM based Similarity Measurement with Spectral Clustering for Speaker Diarization
• SiGAN: Siamese Generative Adversarial Network for Identity-Preserving Face Hallucination
• Adversarial Speaker Verification

Cited by in corpus (3)

• In defence of metric learning for speaker recognition
• NPU Speaker Verification System for INTERSPEECH 2020 Far-Field Speaker Verification Challenge
• Multi-Level Transfer Learning from Near-Field to Far-Field Speaker Verification