Delving into VoxCeleb: environment invariant speaker recognition

Research in speaker recognition has recently seen significant progress due to the application of neural network models and the availability of new large-scale datasets. There has been a plethora of work in search of more powerful architectures or loss functions suitable for the task, but they do not consider what information is learnt by the models aside from being able to predict the given labels. In this work, we introduce an environment adversarial training framework in which the network can effectively learn speaker-discriminative and environment-invariant embeddings without explicit domain shift during training. This allows the network to generalise better in unseen conditions. The method is evaluated on both speaker identification and verification tasks using the VoxCeleb dataset, on which we demonstrate significant performance improvements over baselines.