IEEE - Institute of Electrical and Electronics Engineers, Inc. - Low-rank matrix factorization for Deep Neural Network training with high-dimensional output targets

ICASSP 2013 - IEEE International Conference on Acoustics, Speech and Signal Processing

Author(s): Tara N. Sainath ; Brian Kingsbury ; Vikas Sindhwani ; Ebru Arisoy ; Bhuvana Ramabhadran
Publisher: IEEE - Institute of Electrical and Electronics Engineers, Inc.
Publication Date: 1 May 2013
Conference Location: Vancouver, BC, Canada
Conference Date: 26 May 2013
Page(s): 6,655 - 6,659
ISBN (Electronic): 978-1-4799-0356-6
ISSN (Paper): 1520-6149
DOI: 10.1109/ICASSP.2013.6638949
Regular:

While Deep Neural Networks (DNNs) have achieved tremendous success for large vocabulary continuous speech recognition (LVCSR) tasks, training of these networks is slow. One reason is that DNNs are... View More

Advertisement