Covariance Matrix Enhancement Approach to Train Robust Gaussian Mixture Models of Speech Data

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F13%3A43919379" target="_blank" >RIV/49777513:23520/13:43919379 - isvavai.cz</a>

Výsledek na webu

<a href="http://link.springer.com/chapter/10.1007%2F978-3-319-01931-4_13" target="_blank" >http://link.springer.com/chapter/10.1007%2F978-3-319-01931-4_13</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-319-01931-4_13" target="_blank" >10.1007/978-3-319-01931-4_13</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Covariance Matrix Enhancement Approach to Train Robust Gaussian Mixture Models of Speech Data

Popis výsledku v původním jazyce

An estimation of parameters of a multivariate Gaussian Mixture Model is usually based on a criterion (e.g. Maximum Likelihood) that is focused mostly on training data. Therefore, testing data, which were not seen during the training procedure, may causeproblems. Moreover, numerical instabilities can occur (e.g. for low-occupied Gaussians especially when working with full-covariance matrices in high-dimensional spaces). Another question concerns the number of Gaussians to be trained for a specific dataset. The approach proposed in this paper can handle all these issues. It is based on an assumption that the training and testing data were generated from the same source distribution. The key part of the approach is to use a criterion based on the sourcedistribution rather than using the training data itself. It is shown how to modify an estimation procedure in order to fit the source distribution better (despite the fact that it is unknown), and subsequently new estimation algorithm fo

Název v anglickém jazyce

Covariance Matrix Enhancement Approach to Train Robust Gaussian Mixture Models of Speech Data

Popis výsledku anglicky

An estimation of parameters of a multivariate Gaussian Mixture Model is usually based on a criterion (e.g. Maximum Likelihood) that is focused mostly on training data. Therefore, testing data, which were not seen during the training procedure, may causeproblems. Moreover, numerical instabilities can occur (e.g. for low-occupied Gaussians especially when working with full-covariance matrices in high-dimensional spaces). Another question concerns the number of Gaussians to be trained for a specific dataset. The approach proposed in this paper can handle all these issues. It is based on an assumption that the training and testing data were generated from the same source distribution. The key part of the approach is to use a criterion based on the sourcedistribution rather than using the training data itself. It is shown how to modify an estimation procedure in order to fit the source distribution better (despite the fact that it is unknown), and subsequently new estimation algorithm fo

Druh

D - Stať ve sborníku

CEP obor

JD - Využití počítačů, robotika a její aplikace

OECD FORD obor

—

Projekt

<a href="/cs/project/TA01011264" target="_blank" >TA01011264: Eliminace jazykových bariér handicapovaných diváků České televize II</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2013

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Speech and Computer

ISBN

978-3-319-01930-7

ISSN

0302-9743

e-ISSN

—

Počet stran výsledku

8

Strana od-do

92-99

Název nakladatele

Springer

Místo vydání

Cham

Místo konání akce

Pilzen, Czech Republic

Datum konání akce

1. 9. 2013

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—