Advanced search
1 file | 148.25 KB

Extending reservoir computing with random static projections: a hybrid between extreme learning and RC

Author
Organization
Abstract
Reservoir Computing is a relatively new paradigm in the field of neural networks that has shown promise in applications where traditional recurrent neural networks have performed poorly. The main advantage of using reservoirs is that only the output weights are trained, reducing computational requirements significantly. There is a trade-off, however, between the amount of memory a reservoir can possess and its capability of mapping data into a highly non-linear transformation space. A new, hybrid architecture, combining a reservoir with an extreme learning machine, is presented which overcomes this trade-off, whose performance is demonstrated on a 4th order polynomial modelling task and an isolated spoken digit recognition task.
Keywords
machine learning, neural networks, Reservoir Computing

Downloads

  • Untitled.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 148.25 KB

Citation

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

Chicago
Butcher, John, David Verstraeten, Benjamin Schrauwen, Charles Day, and Peter Haycock. 2010. “Extending Reservoir Computing with Random Static Projections: a Hybrid Between Extreme Learning and RC.” In European Symposium on Artificial Neural Networks, 18th, Proceedings, 303–308. Evere, Belgium: D-Side.
APA
Butcher, J., Verstraeten, D., Schrauwen, B., Day, C., & Haycock, P. (2010). Extending reservoir computing with random static projections: a hybrid between extreme learning and RC. European Symposium on Artificial Neural Networks, 18th, Proceedings (pp. 303–308). Presented at the 18th European Symposium on Artificial Neural Networks (ESANN 2010), Evere, Belgium: D-Side.
Vancouver
1.
Butcher J, Verstraeten D, Schrauwen B, Day C, Haycock P. Extending reservoir computing with random static projections: a hybrid between extreme learning and RC. European Symposium on Artificial Neural Networks, 18th, Proceedings. Evere, Belgium: D-Side; 2010. p. 303–8.
MLA
Butcher, John, David Verstraeten, Benjamin Schrauwen, et al. “Extending Reservoir Computing with Random Static Projections: a Hybrid Between Extreme Learning and RC.” European Symposium on Artificial Neural Networks, 18th, Proceedings. Evere, Belgium: D-Side, 2010. 303–308. Print.
@inproceedings{1024136,
  abstract     = {Reservoir Computing is a relatively new paradigm in the field of neural networks that has shown promise in applications where traditional recurrent neural networks have performed poorly. The main advantage of using reservoirs is that only the output weights are trained, reducing computational requirements significantly. There is a trade-off, however, between the amount of memory a reservoir can possess and its capability of mapping data into a highly non-linear transformation space. A new, hybrid architecture, combining a reservoir with an extreme learning machine, is presented which overcomes this trade-off, whose performance is demonstrated on a 4th order polynomial modelling task and an isolated spoken digit recognition task.},
  author       = {Butcher, John and Verstraeten, David and Schrauwen, Benjamin and Day, Charles and Haycock, Peter},
  booktitle    = {European Symposium on Artificial Neural Networks, 18th, Proceedings},
  keyword      = {machine learning,neural networks,Reservoir Computing},
  language     = {eng},
  location     = {Bruges, Belgium},
  pages        = {303--308},
  publisher    = {D-Side},
  title        = {Extending reservoir computing with random static projections: a hybrid between extreme learning and RC},
  year         = {2010},
}