Advanced search
1 file | 410.69 KB

Continuous digit recognition in noise: reservoirs can do an excellent job!

Author
Organization
Project
ORGANIC (Self-organized recurrent neural learning for language processing)
Abstract
In this paper a formerly proposed continuous digit recognition system based on Reservoir Computing (RC) is improved in two respects: (1)the single reservoir is substituted by a stack of reservoirs, and (2)the straightforward mapping of reservoir outputs to state likelihoods is replaced by a trained non-parametric mapping. Furthermore, it is shown that a reservoir-based method can improve a model trained on clean speech to work better in a noisy condition from which it has a number of unknown digit string recordings available. The first two improvements have lead to a system that outperforms a HMM-based system with the same noise robust features as input. The model adaptation offers a promising supplementary gain at modest noise levels.
Keywords
Noise Robustness, Model Adaptation, Reservoir Computing, Acoustic Modeling

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 410.69 KB

Citation

Please use this url to cite or link to this publication:

Chicago
Jalalvand, Azarakhsh, Fabian Triefenbach, and Jean-Pierre Martens. 2012. “Continuous Digit Recognition in Noise: Reservoirs Can Do an Excellent Job!” In 13th Annual Conference of the International Speech Communication Association, Proceedings, 1-3:1802–1805. International Speech Communication Association (ISCA).
APA
Jalalvand, A., Triefenbach, F., & Martens, J.-P. (2012). Continuous digit recognition in noise: reservoirs can do an excellent job! 13th Annual conference of the International Speech Communication Association, Proceedings (Vol. 1–3, pp. 1802–1805). Presented at the 13th Annual conference of the International Speech Communication Association (Interspeech - 2012), International Speech Communication Association (ISCA).
Vancouver
1.
Jalalvand A, Triefenbach F, Martens J-P. Continuous digit recognition in noise: reservoirs can do an excellent job! 13th Annual conference of the International Speech Communication Association, Proceedings. International Speech Communication Association (ISCA); 2012. p. 1802–5.
MLA
Jalalvand, Azarakhsh, Fabian Triefenbach, and Jean-Pierre Martens. “Continuous Digit Recognition in Noise: Reservoirs Can Do an Excellent Job!” 13th Annual Conference of the International Speech Communication Association, Proceedings. Vol. 1–3. International Speech Communication Association (ISCA), 2012. 1802–1805. Print.
@inproceedings{3054587,
  abstract     = {In this paper a formerly proposed continuous digit recognition system based on Reservoir Computing (RC) is improved in two respects: (1)the single reservoir is substituted by a stack of reservoirs, and (2)the straightforward mapping of reservoir outputs to state likelihoods is replaced by a trained non-parametric mapping. Furthermore, it is shown that a reservoir-based method can improve a model trained on clean speech to work better in a noisy condition from which it has a number of unknown digit string recordings available. The first two improvements have lead to a system that outperforms a HMM-based system with the same noise robust features as input. The model adaptation offers a promising supplementary gain at modest noise levels.},
  articleno    = {644},
  author       = {Jalalvand, Azarakhsh and Triefenbach, Fabian and Martens, Jean-Pierre},
  booktitle    = {13th Annual conference of the International Speech Communication Association, Proceedings},
  isbn         = {9781622767595},
  keyword      = {Noise Robustness,Model Adaptation,Reservoir Computing,Acoustic Modeling},
  language     = {eng},
  location     = {Portland, OR, USA},
  pages        = {644:1802--644:1805},
  publisher    = {International Speech Communication Association (ISCA)},
  title        = {Continuous digit recognition in noise: reservoirs can do an excellent job!},
  volume       = {1-3},
  year         = {2012},
}

Web of Science
Times cited: