Exploiting Eigenposteriors for Semi-Supervised Training of DNN Acoustic Models with Sequence Discrimination.

Pranay Dighe, Afsaneh Asaei, Herv Bourlard

Browse the full Interspeech paper archive.