Information geometry of topology preserving adaptation
We consider adaptation by topologically smooth transformations with applications to environment and speaker adaptation for robust speech recognition. Specifically, the tradeoff between global affine transformations that fail to capture local variation but preserve topology and local class dependent...
Saved in:
Published in | 2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100) Vol. 6; pp. 3743 - 3746 vol.6 |
---|---|
Main Author | |
Format | Conference Proceeding |
Language | English |
Published |
IEEE
2000
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | We consider adaptation by topologically smooth transformations with applications to environment and speaker adaptation for robust speech recognition. Specifically, the tradeoff between global affine transformations that fail to capture local variation but preserve topology and local class dependent bias transformations that have more resolution but may destroy the topology of the reference model is addressed. We cast the problem of topology preservation of the reference model in an information divergence geometry framework and derive a class of alternating minimization algorithms that aims to preserve topology explicitly during adaptation. |
---|---|
ISBN: | 9780780362932 0780362934 |
ISSN: | 1520-6149 2379-190X |
DOI: | 10.1109/ICASSP.2000.860216 |