A hybrid coder for hidden Markov models using a recurrent neural network

Yoshua Bengio, Régis Cardin, Renato De Mori, Yves Normandin

Abstract

A hybrid coder is introduced for obtaining descriptions of speech patterns. This coder uses vector quantization (VQ) techniques on mel-scale cepstral coefficients and their derivatives together with a recurrent network (RN) for describing suprasegmental features of speech. The purpose of these features is to focus the search when hidden Markov models (HMMs) are used for speech unit or word models. Preliminary experiments of speaker-independent connected digit recognition show that using a hybrid coder based on a RN improves recognition performance.<
BibTeX
@inproceedings{bengio1990hybrid,
  author = {Yoshua Bengio and Régis Cardin and Renato De Mori and Yves Normandin},
  title = {{A hybrid coder for hidden Markov models using a recurrent neural network}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {1990},
  doi = {10.1109/icassp.1990.115768},
}

Loading related content...

Figure