On the Determination of Epsilon during Discriminative GMM Training

Author

Guido, Rodrigo Capobianco ; Chen, Shi-Huang ; Junior, Sylvio Barbon ; Souza, Leonardo Mendes ; Vieira, Lucimar Sasso ; Rodrigues, Luciene Cavalcanti ; Escola, Joao Paulo Lemos ; Zulato, Paulo Ricardo Franchi ; Lacerda, Michel Alves ; Ribeiro, Jussara

Author_Institution

SpeechLab, Univ. of Sao Paulo, Sao Carlos, Brazil

fYear

2010

fDate

13-15 Dec. 2010

Firstpage

362

Lastpage

364

Abstract

Discriminative training of Gaussian Mixture Models (GMMs) for speech or speaker recognition purposes is usually based on the gradient descent method, in which the iteration step-size, epsilon, uses to be defined experimentally. In this letter, we derive an equation to adaptively determine epsilon, by showing that the second-order Newton-Raphson iterative method to find roots of equations is equivalent to the gradient descent algorithm.

Keywords

Gaussian processes; Newton-Raphson method; gradient methods; speaker recognition; EPSILON; GMM; Gaussian mixture model; Newton Raphson method; discriminative training; gradient descent algorithm; gradient descent method; iterative method; speaker recognition; speech recognition; Markov Models; discriminative training of Gaussian Mixture Models (GMMs); speaker identification; speech recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Multimedia (ISM), 2010 IEEE International Symposium on

Conference_Location

Taichung

Print_ISBN

978-1-4244-8672-4

Electronic_ISBN

978-0-7695-4217-1

Type

conf

DOI

10.1109/ISM.2010.66

Filename

5693868