Coupling a Generative Model With a Discriminative Learning Framework for Speaker Verification | IEEE Journals & Magazine | IEEE Xplore