Tandem connectionist feature extraction for conventional HMM systems
Cached
Download Links
| Citations: | 88 - 15 self |
BibTeX
@MISC{Hermansky_tandemconnectionist,
author = {Hynek Hermansky and Daniel P. W. Ellis and Sangita Sharma},
title = {Tandem connectionist feature extraction for conventional HMM systems},
year = {}
}
Years of Citing Articles
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
Hidden Markov model speech recognition systems typically use Gaussian mixture models to estimate the distributions of decorrelated acoustic feature vectors that correspond to individual subword units. By contrast, hybrid connectionist-HMM systems use discriminatively-trained neural networks to estimate the probability distribution among subword units given the acoustic observations. In thisworkweshowalargeimprovementinwordrecognitionperformancebycombiningneural -netdiscriminativefeature processingwithGaussian-mixturedistributionmodeling.Bytrainingthenetworktogeneratethesubwordprobabilityposteriors, thenusingtransformationsoftheseestimatesasthebasefeatures foraconventionally-trainedGaussian-mixturebasedsystem,we achieverelativeerrorratereductions of 35% or mor eonthemulticondition Aurora noisy continuous digits task.
