Tied Variational Autoencoder Backends for i-Vector Speaker Recognition.

Jess Villalba, Niko Brmmer, Najim Dehak

Browse the full Interspeech paper archive.