Experiments with time delay networks and dynamic time warping for speaker independent isolated digits recognition

We descrtbe in thls paper a speaker Independent. global word recognition task ustng time delay networks. We flrst descrtbe these networks as a way for leamlng feature extractors by constralned back-propagatlon. Such a tlme-delay network ls shown to be capable of deallng wlth a test task: French dlgtt recognttlon. The results are dlscussed and compared, on the same data sets, wlth those obtained systems on test

[1]  J.-L. Gauvain,et al.  A dynamic programming processor for speech recognition , 1988, Proceedings of the IEEE 1988 Custom Integrated Circuits Conference.

[2]  Alex Waibel,et al.  Phoneme recognition: neural networks vs. hidden Markov models vs. hidden Markov models , 1988, ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing.

[3]  D Zipser,et al.  Learning the hidden structure of speech. , 1988, The Journal of the Acoustical Society of America.

[4]  Patrick Gallinari,et al.  EVALUATION OF NETWORK ARCHITECTURES ON TEST LEARNING TASKS. , 1987 .

[5]  J.-L. Gauvain,et al.  A syllable-based isolated word recognition experiment , 1986, ICASSP '86. IEEE International Conference on Acoustics, Speech, and Signal Processing.

[6]  T. D. Harrison,et al.  Boltzmann machines for speech recognition , 1986 .

[7]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[8]  S.E. Levinson,et al.  Structural methods in automatic speech recognition , 1985, Proceedings of the IEEE.