DocumentCode :
542179
Title :
Minimum Phone Error and I-smoothing for improved discriminative training
Author :
Povey, D. ; Woodland, P.C.
Author_Institution :
Cambridge University Engineering Dept, Trumpington St., CB2 1PZ U.K.
Volume :
1
fYear :
2002
fDate :
13-17 May 2002
Abstract :
In this paper we introduce the Minimum Phone Error (MPE) and Minimum Word Error (MWE) criteria for the discriminative training of HMM systems. The MPE/MWE criteria are smoothed approximations to the phone or word error rate respectively. We also discuss I-smoothing which is a novel technique for smoothing discriminative training criteria using statistics for maximum likelihood estimation (MLE). Experiments have been performed on the Switchboard/Call Home corpora of telephone conversations with up to 265 hours of training data. It is shown that for the maximum mutual information estimation (MMIE) criterion, I-smoothing reduces the word error rate (WER) by 0.4% absolute over the MMIE baseline. The combination of MPE and I-smoothing gives an improvement of 1 % over MMIE and a total reduction in WER of 4.8% absolute over the original MLE system.
Keywords :
Decoding; Hidden Markov models; Maximum likelihood estimation; Switches; Training; Variable speed drives; Vocabulary;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Acoustics, Speech, and Signal Processing (ICASSP), 2002 IEEE International Conference on
Conference_Location :
Orlando, FL, USA
ISSN :
1520-6149
Print_ISBN :
0-7803-7402-9
Type :
conf
DOI :
10.1109/ICASSP.2002.5743665
Filename :
5743665
Link To Document :
بازگشت