Detail publikace

A Symmetrization of the Subspace Gaussian Mixture Model

Originální název

A Symmetrization of the Subspace Gaussian Mixture Model

Anglický název

A Symmetrization of the Subspace Gaussian Mixture Model

Jazyk

en

Originální abstrakt

We have described a modification to the Subspace Gaussian Mixture Model which we call the Symmetric SGMM. This is a very natural extension which removes an asymmetry in the way the Gaussian mixture weights were previously computed. The extra computation is minimal but the memory used for the acoustic model is nearly doubled. Our experimental results were inconsistent: on one setup we got a large improvement of 1.5% absolute, and on another setup it was much smaller.

Anglický abstrakt

We have described a modification to the Subspace Gaussian Mixture Model which we call the Symmetric SGMM. This is a very natural extension which removes an asymmetry in the way the Gaussian mixture weights were previously computed. The extra computation is minimal but the memory used for the acoustic model is nearly doubled. Our experimental results were inconsistent: on one setup we got a large improvement of 1.5% absolute, and on another setup it was much smaller.

BibTex


@inproceedings{BUT76375,
  author="Daniel {Povey} and Martin {Karafiát} and Arnab {Ghoshal} and Petr {Schwarz}",
  title="A Symmetrization of the Subspace Gaussian Mixture Model",
  annote="We have described a modification to the Subspace Gaussian Mixture Model which we
call the Symmetric SGMM. This is a very natural extension which removes an
asymmetry in the way the Gaussian mixture weights were previously computed. The
extra computation is minimal but the memory used for the acoustic model is nearly
doubled. Our experimental results were inconsistent: on one setup we got a large
improvement of 1.5% absolute, and on another setup it was much smaller.",
  address="IEEE Signal Processing Society",
  booktitle="Proceedings of 2011 IEEE International Conference on Acoustics, Speech, and Signal Processing",
  chapter="76375",
  edition="NEUVEDEN",
  howpublished="print",
  institution="IEEE Signal Processing Society",
  year="2011",
  month="may",
  pages="4504--4507",
  publisher="IEEE Signal Processing Society",
  type="conference paper"
}