Discriminative Feature Extraction Based on Sequential Variational Autoencoder for Speaker Recognition | IEEE Conference Publication | IEEE Xplore