Cosine-Distance Virtual Adversarial Training for Semi-Supervised Speaker-Discriminative Acoustic Embeddings.

Florian L. Kreyssig, Philip C. Woodland

Browse the full Interspeech paper archive.