Advanced search
1 file | 522.24 KB

Feature enhancement with a reservoir-based denoising auto encoder

Azarakhsh Jalalvand (UGent) , Kris Demuynck (UGent) and Jean-Pierre Martens (UGent)
Author
Organization
Abstract
Recently, automatic speech recognition has advanced significantly by the introduction of deep neural networks for acoustic modeling. However, there is no clear evidence yet that this does not come at the price of less generalization to conditions that were not present during training. On the other hand, acoustic modeling with Reservoir Computing (RC) did not offer improved clean speech recognition but it leads to good robustness against noise and channel distortions. In this paper, the aim is to establish whether adding feature denoising in the front-end can further improve the robustness of an RC-based recognizer, and if so, whether one can devise an RC-based Denoising Auto Encoder that outperforms a traditional denoiser like the ETSI Advanced Front-End. In order to answer these questions, experiments are conducted on the Aurora-2 benchmark.
Keywords
reservoir computing, robust speech recognition, recurrent neural networks, denoising auto encoder, ROBUST SPEECH RECOGNITION, NOISE

Downloads

  • PID2941323.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 522.24 KB

Citation

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

Chicago
Jalalvand, Azarakhsh, Kris Demuynck, and Jean-Pierre Martens. 2013. “Feature Enhancement with a Reservoir-based Denoising Auto Encoder.” In IEEE International Symposium on Signal Processing and Information Technology, 227–232. New York, NY, USA: IEEE.
APA
Jalalvand, A., Demuynck, K., & Martens, J.-P. (2013). Feature enhancement with a reservoir-based denoising auto encoder. IEEE International Symposium on Signal Processing and Information Technology (pp. 227–232). Presented at the 2013 IEEE International symposium on Signal Processing and Information Technology (IEEE ISSPIT 2013), New York, NY, USA: IEEE.
Vancouver
1.
Jalalvand A, Demuynck K, Martens J-P. Feature enhancement with a reservoir-based denoising auto encoder. IEEE International Symposium on Signal Processing and Information Technology. New York, NY, USA: IEEE; 2013. p. 227–32.
MLA
Jalalvand, Azarakhsh, Kris Demuynck, and Jean-Pierre Martens. “Feature Enhancement with a Reservoir-based Denoising Auto Encoder.” IEEE International Symposium on Signal Processing and Information Technology. New York, NY, USA: IEEE, 2013. 227–232. Print.
@inproceedings{4214421,
  abstract     = {Recently, automatic speech recognition has advanced significantly by the introduction of deep neural networks for acoustic modeling. However, there is no clear evidence yet that this does not come at the price of less generalization to conditions that were not present during training. On the other hand, acoustic modeling with Reservoir Computing (RC) did not offer improved clean speech recognition but it leads to good robustness against noise and channel distortions. In this paper, the aim is to establish whether adding feature denoising in the front-end can further improve the robustness of an RC-based recognizer, and if so, whether one can devise an RC-based Denoising Auto Encoder that outperforms a traditional denoiser like the ETSI Advanced Front-End. In order to answer these questions, experiments are conducted on the Aurora-2 benchmark.},
  author       = {Jalalvand, Azarakhsh and Demuynck, Kris and Martens, Jean-Pierre},
  booktitle    = {IEEE International Symposium on Signal Processing and Information Technology},
  isbn         = {9781479947966},
  issn         = {2162-7843},
  language     = {eng},
  location     = {Athens, Greece},
  pages        = {227--232},
  publisher    = {IEEE},
  title        = {Feature enhancement with a reservoir-based denoising auto encoder},
  url          = {http://dx.doi.org/10.1109/ISSPIT.2013.6781884},
  year         = {2013},
}

Altmetric
View in Altmetric
Web of Science
Times cited: