Loading...
conference paper
Subspace Gaussian Mixture Models for speech recognition
2010
2010 IEEE International Conference on Acoustics, Speech and Signal Processing
We describe an acoustic modeling approach in which all phonetic states share a common Gaussian Mixture Model structure, and the means and mixture weights vary in a subspace of the total parameter space. We call this a Subspace Gaussian Mixture Model (SGMM). Globally shared parameters define the subspace. This style of acoustic model allows for a much more compact representation and gives better results than a conventional modeling approach, particularly with smaller amounts of training data.
Type
conference paper
Authors
Povey, Daniel
•
Burget, Lukas
•
Agarwal, Mohit
•
•
Feng, Kai
•
Ghoshal, Arnab
•
Glembek, Ondrej
•
Goel, Nagendra Kumar
•
Karafiat, Martin
•
Rastrow, Ariya
Publication date
2010
Published in
2010 IEEE International Conference on Acoustics, Speech and Signal Processing
Start page
4330
End page
4333
Peer reviewed
REVIEWED
EPFL units
Event name | Event date |
2010 | |
Available on Infoscience
November 19, 2014
Use this identifier to reference this record