Name
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. |
Year | DOI | Venue |
|---|---|---|
2010 | 10.1109/ISM.2010.66 | ISM |
Keywords | Field | DocType |
second-order newton-raphson iterative method,discriminative gmm training,gradient descent algorithm,gaussian mixture models,iteration step-size,gradient descent method,speaker recognition purpose,discriminative training,markov processes,speech processing,speech,gmm,gaussian distribution,gradient descent,iteration method,markov models,second order,newton raphson method,markov model,speech recognition,loudspeakers,acoustics,gaussian mixture model,gaussian processes,iterative method,speaker recognition,iterative methods,mathematical model,newton raphson | Speech processing,Gradient descent,Pattern recognition,Markov model,Computer science,Iterative method,Speaker recognition,Gaussian process,Artificial intelligence,Discriminative model,Mixture model | Conference |
ISBN | Citations | PageRank |
978-0-7695-4217-1 | 0 | 0.34 |
References | Authors | |
6 | 11 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Rodrigo Capobianco Guido | 1 | 161 | 27.59 |
| Shi-Huang Chen | 2 | 12 | 3.05 |
| Sylvio Barbon Júnior | 3 | 50 | 14.05 |
| Leonardo Mendes Souza | 4 | 1 | 0.72 |
| Lucimar Sasso Vieira | 5 | 20 | 4.85 |
| Luciene Cavalcanti Rodrigues | 6 | 3 | 1.12 |
| Joao Paulo Lemos Escola | 7 | 0 | 0.34 |
| Paulo Ricardo Franchi Zulato | 8 | 1 | 0.73 |
| Michel Alves Lacerda | 9 | 1 | 0.73 |
| Jussara Ribeiro | 10 | 1 | 0.73 |
| Barbon Junior, S. | 11 | 0 | 0.68 |