Text-to-speech inspired duration modeling for improved whole-word acoustic models.

Keith Kintzley, Aren Jansen, Hynek Hermansky

Browse the full Interspeech paper archive.