Wise teachers train better DNN acoustic models.

Ryan Price Ken-ichi Iso Koichi Shinoda

Published in: EURASIP J. Audio Speech Music. Process. (2016)

Keyphrases

acoustic models
speech recognition
learning environment
training process
hidden markov models
learning process
pairwise
e learning
automatic speech recognition
neural network
pattern recognition
speech recognizer
broadcast news
metadata
image processing
discriminative training