Neural Networks, Learning Automata and Iterated Function Systems

  • P. C. Bressloff
  • J. Stark

Abstract

An overview is given of certain underlying relationships between neural networks and Iterated Function Systems. Possible applications to data compression and stochastic learning automata are discussed.

Keywords

Convolution Mist 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [Barn85]
    Barnsley, M.F., Demko, S., Elton, J., and Geronimo, J., Markov processes arising from iteration with place-dependent probabilities, Pre-print, Georgia Inst. Technology, 1985.Google Scholar
  2. [Barn88a]
    Barnsley, M.F., Fractals Everywhere, San Diego: Academic Press, 1988.MATHGoogle Scholar
  3. [Barn88b]
    Barnsley, M.F., and Sloan, A.D., A better way to compress images, Byte, Vol. 13, pp. 215–223, 1988.Google Scholar
  4. [Bart81]
    Barto, A.G., Sutton, R.S., and Brouwer, P.S., Associative search network, Biol. Cybern., Vol. 40, pp. 201–211, 1981.MATHCrossRefGoogle Scholar
  5. [Bart85]
    Barto, A.G., and Anandan, P., Pattern-recognising stochastic learning automata, IEEE Trans. Syst. Man and Cybrn., Vol. 15, pp. 360–375, 1985.MathSciNetMATHGoogle Scholar
  6. [Bres89a]
    Bressloff, P.C., Neural networks and random iterative maps, in Neural Computation, Mannion, C, and Taylor, J.G., Eds., pp. 155–162, Bristol, UK: Adam Hilger, 1989.Google Scholar
  7. [Bres89b]
    Bressloff, P.C., and Stark, J., Associative reinforcement learning based on iterated function systems, submitted to IEEE Trans. Syst. Man and Cybern., 1989.Google Scholar
  8. [Bres90]
    Bressloff, P.C., and Taylor, J.G., Random iterative networks, Phys. Rev. A, Vol. 41, pp. 1126–1137, 1990.MathSciNetCrossRefGoogle Scholar
  9. [Elto87]
    Elton, J.H., An ergodic theorem for iterated maps, Ergod. Th. and Dynam. Sys., Vol. 7, pp. 481–488, 1987.MathSciNetMATHCrossRefGoogle Scholar
  10. [Erdo39]
    Erdos, P., On a family of symmetric Bernouilli convolutions, Amer. Jour. Math., Vol. 61, pp. 974–976, 1939.CrossRefGoogle Scholar
  11. [Fede69]
    Federer, H., Geometric Measure Theory, New York: Springer-Verlag, 1969.MATHGoogle Scholar
  12. [Falc85]
    Falconer, K.J., The Geometry of Fractal Sets, Cambridge, UK: Cambridge Univ. Press, 1985.MATHCrossRefGoogle Scholar
  13. [Hutc81]
    Hutchinson, J.F., Fractals and self-similarity,Indiana Univ. Jour, of Math., Vol. 30, pp. 713–747, 1981.MathSciNetMATHCrossRefGoogle Scholar
  14. [Karl53]
    Karlin, S., Some random walks arising in learning models, Pacific Jour. Math., Vol. 3, pp. 725–756, 1953.MathSciNetMATHGoogle Scholar
  15. [Laks81]
    Lakshmivarahan, S., Learning Algorithms Theory and Applications, New York: Springer-Verlag, 1981.MATHGoogle Scholar
  16. [Nare74]
    Narendra, K.S., and Thathachar, M.A.L., Learning automata—A survey, IEEE, Trans. Syst. Man and Cybern., Vol. 4, pp. 323–334, 1974.MathSciNetMATHCrossRefGoogle Scholar
  17. [Norm68]
    Norman, M.F., Some convergence theorems for stochastic learning models with distance diminishing operators, Jour. Math. Psychol, Vol. 5, pp. 61–101, 1968.MathSciNetMATHCrossRefGoogle Scholar
  18. [Norm72]
    Norman, M.F., Markov Processes and Learning Models, New York: Academic Press, 1972.MATHGoogle Scholar
  19. [Prus88]
    Prusinkiewicz, P., and Sandness, G., Koch curves as attractors and repellors, IEEE Comput. Graph, and Appl, Vol. 8, pp. 26–40, 1988.Google Scholar
  20. [Redd88]
    Reddaway, S.F., Fractal graphics and image compression on a DAP, in International Specialist Seminar on the Design and Application of Parallel Digital Processors(IEE Conf. Publ. 298), p. 201, 1988.Google Scholar
  21. [Sene81]
    Seneta, W., Non-negative Matrices and Markov Chains, New York: Springer-Verlag, 1981.MATHGoogle Scholar
  22. [Star89]
    Stark, J., Iterated function systems as neural networks, submitted to Neural Networks, 1989.Google Scholar

Copyright information

© Springer-Verlag New York Inc. 1991

Authors and Affiliations

  • P. C. Bressloff
  • J. Stark

There are no affiliations available

Personalised recommendations