An improved MMIE training algorithm for speaker-independent, small vocabulary, continuous speech recognition
作者： Y. Normandin , S.D. Morgera
DOI: 10.1109/ICASSP.1991.150395
关键词: Estimation theory 、 Word error rate 、 Markov process 、 NIST 、 Mutual information 、 Algorithm 、 Gaussian 、 Information theory 、 Rate of convergence 、 Computer science 、 Speech recognition 、 Hidden Markov model 、 Vocabulary
摘要: Recently, Gopalakrishnan et al. (1989) introduced a reestimation formula for discrete HMMs (hidden Markov models) which applies to rational objective functions like the MMIE (maximum mutual information estimation) criterion. The authors analyze and show how its convergence rate can be substantially improved. They introduce corrective training algorithm, which, when applied TI/NIST connected digit database, has made it possible reduce string error by close 50%. Gopalakrishnan's result is extended continuous case proposing new estimating mean variance parameters of diagonal Gaussian densities. >
computer.org
sci-hub.se 参考文章(8)
P.S. Gopalakrishnan, D. Kanevsky, A. Nadas, D. Nahamoo, A generalization of the Baum algorithm to rational objective functions international conference on acoustics, speech, and signal processing. pp. 631- 634 ,(1989) , 10.1109/ICASSP.1989.266506
B. Merialdo, Phonetic recognition using hidden Markov models and maximum mutual information training international conference on acoustics speech and signal processing. pp. 111- 114 ,(1988) , 10.1109/ICASSP.1988.196524
L. Bahl, P. Brown, P. de Souza, R. Mercer, Maximum mutual information estimation of hidden Markov model parameters for speech recognition international conference on acoustics, speech, and signal processing. ,vol. 11, pp. 49- 52 ,(1986) , 10.1109/ICASSP.1986.1169179
A. Nadas, A decision theorectic formulation of a training problem in speech recognition and a comparison of training by unconditional versus conditional maximum likelihood IEEE Transactions on Acoustics, Speech, and Signal Processing. ,vol. 31, pp. 814- 817 ,(1983) , 10.1109/TASSP.1983.1164173
R. Cardin, Y. Normandin, R. De Mori, High performance connected digit recognition using maximum mutual information estimation international conference on acoustics, speech, and signal processing. pp. 533- 536 ,(1991) , 10.1109/ICASSP.1991.150394
Leonard E. Baum, J. A. Eagon, An inequality with applications to statistical estimation for probabilistic functions of Markov processes and to a model for ecology Bulletin of the American Mathematical Society. ,vol. 73, pp. 360- 363 ,(1967) , 10.1090/S0002-9904-1967-11751-8
A. Nadas, D. Nahamoo, M.A. Picheny, On a model-robust training method for speech recognition IEEE Transactions on Acoustics, Speech, and Signal Processing. ,vol. 36, pp. 1432- 1436 ,(1988) , 10.1109/29.90371
Y.-L. Chow, Maximum mutual information estimation of HMM parameters for continuous speech recognition using the N-best algorithm international conference on acoustics, speech, and signal processing. pp. 701- 704 ,(1990) , 10.1109/ICASSP.1990.115863