A Dynamical Architecture for a Radial Basis Function Network

  • Bernard Lemarié
  • Anne-Gaelle Debroise


We present a dynamical architecture for a Radial Basis Function Network. The scheme is based on the Simulated Annealing procedure for learning. Increase of performances with respect to classical methods and opportunity to vary the size of the network are reported.


Mean Square Error Simulated Annealing Learning Phase Radial Basis Function Network Bayesian Probability 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    AArts E., Korst J.: Simulated annealing and Boltzmann Machine. John Wiley & Sons, (1989).Google Scholar
  2. 2.
    Buhman J, Kuhnel H.: Neural Comp., 5, 75, (1993).CrossRefGoogle Scholar
  3. 3.
    Hampshire J. B., Pearlmutter B.: Proceedings of the 1990 connectionnist models summer school, Morgan Kaufmann, 159, (1990).Google Scholar
  4. 4.
    Kirkpatrick S., Gelatt C., Vecchi M.: Science 220, 671, (1983).MathSciNetMATHCrossRefGoogle Scholar
  5. 5.
    Lemarié B.: IJCNN’93, 331, (1993).Google Scholar
  6. 6.
    Moody J., Darken C. J.: Neural Comp., 1, 281, (1989).CrossRefGoogle Scholar
  7. 7.
    Ng, Lipmann R.P.: NIPS 3, (1991).Google Scholar
  8. 8.
    Poggio T., Girosi F.: Proceedings of the IEEE, 78, 9, (1990).CrossRefGoogle Scholar
  9. 9.
    Rose K., Gurewitz E., Fox G. C.: Physic. Rev. Let., 65, 945, (1990)CrossRefGoogle Scholar
  10. 10.
    Zang Q.: IRISA, internal report No 709, France, (1993).Google Scholar

Copyright information

© Springer-Verlag/Wien 1995

Authors and Affiliations

  • Bernard Lemarié
    • 1
  • Anne-Gaelle Debroise
    • 1
  1. 1.Service de recherche Technique de la Poste, (SRTP)Nantes CedexFrance

Personalised recommendations