Speaker adaptation of deep neural networks using a hierarchy of output layers.

Ryan Price Ken-ichi Iso Koichi Shinoda

Published in: SLT (2014)

Keyphrases

neural network
multi layer
speaker adaptation
multiple layers
single layer
pattern recognition
maximum likelihood
artificial neural networks
back propagation
deep learning
speech recognition
deep belief networks
activation function
noisy images
mixture model
multi modal
machine learning