Skip to main content
Log in

Martingale approach to neural networks with hierarchically structured information

  • Published:
Zeitschrift für Physik B Condensed Matter

Abstract

A modified version of a model with hierarchically structured information, originally proposed by Parga and Virasoro, is studied in the context of martingale theory. The number of generations is allowed to be arbitrary but fixed. In the calculation of the storage capacity both the patterns and all their ancestors have to be taken into account as retrieval states. Due to the martingale property, the free energy greatly simplifies and for the retrieval states is shown to reduce to the one of the Hopfield model. So the storage capacity and retrieval quality of both models are identical. Numerical stimulations confirm the theoretical predictions.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. Hopfield, J.J.: Proc. Natl. Acad. Sci. USA79, 2554 (1982);81 3088 (1984)

    Google Scholar 

  2. Hebb, D.O.: The Organization of behavior. New York: Wiley 1949

    Google Scholar 

  3. Dotsenko, V.S.: J. Phys. C18 L1017 (1985); Physica140A, 410 (1986)

    Google Scholar 

  4. Feigelman, M.V., Ioffe, L.B.: (Preprint), Landau Institute

  5. Gutfreund, H.: Phys. Rev. A: (to be published)

  6. Parga, N., Virasoro, M.A.: J. Phys. (Paris)47, 1857 (1986)

    Google Scholar 

  7. van Hemmen, J.L.: Phys. Rep.65, 43 (1980), esp. Sect. 1

    Google Scholar 

  8. Cortes, C., Krogh, A., Hertz, J.A.: J. Phys. A: Math. Gen.20, 4449 (1987)

    Google Scholar 

  9. Lamperti, J.: Stochastic processes, App. 2. Berlin, Heidelberg, New York: Springer 1977

    Google Scholar 

  10. Doob, J.: Am. Math. Month.78, 451 (1971)

    Google Scholar 

  11. Chung, K.L.: A course in probability theory, 2nd edn; Chap. 9. New York: Academic Press 1974

    Google Scholar 

  12. Lamperti, J.: Probability, Sect. 7. New York: Benjamin 1966

    Google Scholar 

  13. Amit, D.J., Gutfreund, H., Sompolinsky, H.: Phys. Rev. Lett.55, 1530 (1985); Ann. Phys. (N.Y.)173, 30 (1987)

    Google Scholar 

  14. Weisbuch, G., Fogelman-Soulié, F.: J. Phys. Lett. (Paris)46, L623 (1985)

    Google Scholar 

  15. Mc Eliece, R.J., Posner, E.C., Rodemich, E.R., Venkatesh, S.S.: Cal. Tech.: (Preprint)

  16. van Hemmen, J.L., Palmer, R.G.: J. Phys. A.: Math. Gen.12, 563 (1979)

    Google Scholar 

  17. van Hemmen, J.L., Zagrebnov, V.A.: J. Phys. A.: Math. Gen.20, 3989 (1987)

    Google Scholar 

  18. van Hemmen, J.L.: Phys. Rev. A36, 1959 (1987)

    Google Scholar 

  19. Shiryayev, A.N.: Probability. Berlin, Heidelberg, New York: Springer 1984

    Google Scholar 

  20. Grensing, D., Kühn, R., van Hemmen, J.L.: J. Phys. A: Math. Gen.20, 2935 (1987)

    Google Scholar 

  21. Kinzel, W.: Z. Phys. B-Condensed Matter60, 205 (1985)

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Rights and permissions

Reprints and permissions

About this article

Cite this article

Bös, S., Kühn, R. & van Hemmen, J.L. Martingale approach to neural networks with hierarchically structured information. Z. Physik B - Condensed Matter 71, 261–271 (1988). https://doi.org/10.1007/BF01312798

Download citation

  • Received:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF01312798

Keywords

Navigation