Home

Tandem connectionist feature stream extraction for conventional HMM systems

Hynek Hermansky; Daniel P. W. Ellis; Sangita Sharma

Title:
Tandem connectionist feature stream extraction for conventional HMM systems
Author(s):
Hermansky, Hynek
Ellis, Daniel P. W.
Sharma, Sangita
Date:
Type:
Articles
Department:
Electrical Engineering
Permanent URL:
Book/Journal Title:
2000 IEEE International Conference on Acoustics, Speech, and Signal Processing: Proceedings, 5-9 June, 2000, Hilton Hotel and Convention Center, Istanbul, Turkey
Publisher:
IEEE
Publisher Location:
Piscataway, N.J.
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 this work we show a large improvement in word recognition performance by combining neural-net discriminative feature processing with Gaussian-mixture distribution modeling. By training the network to generate the subword probability posteriors, then using transformations of these estimates as the base features for a conventionally-trained Gaussian-mixture based system, we achieve relative error rate reductions of 35% or more on the multicondition Aurora noisy continuous digits task
Subject(s):
Electrical engineering
Applied mathematics
Publisher DOI:
10.1109/ICASSP.2000.862024
Item views:
39
Metadata:
View

In Partnership with the Center for Digital Research and Scholarship at Columbia University Libraries/Information Services.