Large vocabulary continuous speech recognition with context-dependent DBN-HMMS

The context-independent deep belief network (DBN) hidden Markov model (HMM) hybrid architecture has recently achieved promising results for phone recognition. In this work, we propose a context-dependent DBN-HMM system that dramatically outperforms strong Gaussian mixture model (GMM)-HMM baselines o...

Full description

Saved in:
Bibliographic Details
Published in2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 4688 - 4691
Main Authors Dahl, George E., Dong Yu, Li Deng, Acero, Alex
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.05.2011
Subjects
Accuracy
Acoustics
Artificial neural networks
context-dependent phone
DBN-HMM
deep belief network
Hidden Markov models
LVCSR
Speech recognition
Training
Vocabulary
Online AccessGet full text

Cover

More Information
Summary:The context-independent deep belief network (DBN) hidden Markov model (HMM) hybrid architecture has recently achieved promising results for phone recognition. In this work, we propose a context-dependent DBN-HMM system that dramatically outperforms strong Gaussian mixture model (GMM)-HMM baselines on a challenging, large vocabulary, spontaneous speech recognition dataset from the Bing mobile voice search task. Our system achieves absolute sentence accuracy improvements of 5.8% and 9.2% over GMM-HMMs trained using the minimum phone error rate (MPE) and maximum likelihood (ML) criteria, respectively, which translate to relative error reductions of 16.0% and 23.2%.
ISBN:9781457705380
1457705389
ISSN:1520-6149
DOI:10.1109/ICASSP.2011.5947401