Subspace Gaussian Mixture Models for speech recognition
Daniel Povey, Lukáš Burget, Mohit Agarwal, Pinar Akyazi, Kai Feng, Arnab Kumar Ghoshal, Ondřej Glembek, Nagendra Kumar Goel, Martin Karafiát, Ariya Rastrow, Richard Cameron Rose, Petr Schwarz, Samuel Thomas · 2010
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.