On the Determination of Epsilon during Discriminative GMM Training

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, b...

Full description

Saved in:
Bibliographic Details
Main Authors: Guido, R C, Chen, S, Barbon Junior, Sylvio, Souza, L M, Vieira, L S, Rodrigues, L C, Escola, J P L, Zulato, P R F, Lacerda, M A, Ribeiro, J
Format: Conference Proceeding
Language:English
Subjects:
Acoustics
discriminative training of Gaussian Mixture Models (GMMs)
Equations
Markov Models
Mathematical model
speaker identification
Speech
Speech processing
Speech recognition
Training
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary: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.
DOI:10.1109/ISM.2010.66