DocumentCode :
1368556
Title :
Discriminative weighting of HMM state-likelihoods using the GPD method
Author :
Kwon, O.W. ; Un, C.K.
Author_Institution :
Commun. Res. Lab., Korea Adv. Inst. of Sci. & Technol., Seoul, South Korea
Volume :
3
Issue :
9
fYear :
1996
Firstpage :
257
Lastpage :
259
Abstract :
We propose a new method of finding discriminative state-weights recursively using the generalized probabilistic descent method. This method is implemented with minor modification of the conventional parameter estimation and recognition algorithms by constraining the sum of the state-weights to the number of states in a recognition unit, and can be applied to continuous speech recognition as well as isolated word recognition. We confirm the validity of the method with phoneme-based and word-based state-weighting schemes for three kinds of recognition tasks.
Keywords :
hidden Markov models; parameter estimation; recursive estimation; speech recognition; state estimation; GPD method; HMM state-likelihoods; continuous speech recognition; discriminative weighting; generalized probabilistic descent method; isolated word recognition; parameter estimation; phoneme-based state-weighting; word-based state-weighting; Hidden Markov models; Maximum likelihood estimation; Nonhomogeneous media; Parameter estimation; Probability density function; Recursive estimation; Speech recognition; State estimation; Training data; Viterbi algorithm;
fLanguage :
English
Journal_Title :
Signal Processing Letters, IEEE
Publisher :
ieee
ISSN :
1070-9908
Type :
jour
DOI :
10.1109/97.536594
Filename :
536594
Link To Document :
بازگشت