Show simple item record

Proc. Interspeech 2008

dc.contributor.authorBell, Peter
dc.contributor.authorKing, Simon
dc.date.accessioned2010-10-05T11:49:24Z
dc.date.available2010-10-05T11:49:24Z
dc.date.issued2008en
dc.identifier.urihttp://hdl.handle.net/1842/3839
dc.description.abstractWe consider the problem of parameter estimation in full-covariance Gaussian mixture systems for automatic speech recognition. Due to the high dimensionality of the acoustic feature vector, the standard sample covariance matrix has a high variance and is often poorly-conditioned when the amount of training data is limited. We explain how the use of a shrinkage estimator can solve these problems, and derive a formula for the optimal shrinkage intensity. We present results of experiments on a phone recognition task, showing that the estimator gives a performance improvement over a standard full-covariance systemen
dc.titleA Shrinkage Estimator for Speech Recognition with Full Covariance HMMsen
dc.typeConference Paperen
rps.titleProc. Interspeech 2008en
dc.date.updated2010-10-05T11:49:24Z
dc.date.openingDate2008-09


Files in this item

This item appears in the following Collection(s)

Show simple item record