DocumentCode :
1064941
Title :
Genones: generalized mixture tying in continuous hidden Markov model-based speech recognizers
Author :
Digalakis, Vassilios V. ; Monaco, Peter ; Murveit, Hy
Author_Institution :
Speech Technol. & Res. Lab., SRI Int., Menlo Park, CA, USA
Volume :
4
Issue :
4
fYear :
1996
fDate :
7/1/1996 12:00:00 AM
Firstpage :
281
Lastpage :
289
Abstract :
An algorithm is proposed that achieves a good tradeoff between modeling resolution and robustness by using a new, general scheme for tying of mixture components in continuous mixture-density hidden Markov model (HMM)-based speech recognizers. The sets of HMM states that share the same mixture components are determined automatically using agglomerative clustering techniques. Experimental results on ARPA´s Wall Street Journal corpus show that this scheme reduces errors by 25% over typical tied-mixture systems. New fast algorithms for computing Gaussian likelihoods-the-most time-consuming aspect of continuous-density HMM systems-are also presented. These new algorithms-significantly reduce the number of Gaussian densities that are evaluated with little or no impact on speech recognition accuracy
Keywords :
Gaussian processes; decoding; hidden Markov models; speech coding; speech recognition; ARPA´s Wall Street Journal corpus; Gaussian densities; Gaussian likelihoods; HMM states; agglomerative clustering techniques; algorithm; continuous hidden Markov model-based speech recognizers; decoding; generalized mixture tying; genones; mixture components; modeling resolution; robustness; speech recognition accuracy; Associate members; Automatic speech recognition; Clustering algorithms; Gaussian processes; Hidden Markov models; Robustness; Speech recognition; State estimation; Stochastic processes; Training data;
fLanguage :
English
Journal_Title :
Speech and Audio Processing, IEEE Transactions on
Publisher :
ieee
ISSN :
1063-6676
Type :
jour
DOI :
10.1109/89.506931
Filename :
506931
Link To Document :
بازگشت