Inconsistency between training and testing criteria is a drawback of the hybrid artifcial neural network and hidden Markov model (ANN/HMM) approach to speech recognition. This paper presents an effective method to address this problem by modifying the feedforward neural network training paradigm. Word errors are explicitly incorporated in the training procedure to achieve improved word recognition accuracy. Experiments on a continuous digit database show a reduction in word error rate of more than 17% using the proposed method.
Cite as: Tu, X., Yan, Y., Cole, R. (1997) Matching training and testing criteria in hybrid speech recognition systems. Proc. 5th European Conference on Speech Communication and Technology (Eurospeech 1997), 1943-1946, doi: 10.21437/Eurospeech.1997-516
@inproceedings{tu97_eurospeech,
author={Xin Tu and Yonghong Yan and Ron Cole},
title={{Matching training and testing criteria in hybrid speech recognition systems}},
year=1997,
booktitle={Proc. 5th European Conference on Speech Communication and Technology (Eurospeech 1997)},
pages={1943--1946},
doi={10.21437/Eurospeech.1997-516}
}