Layered Networks for Unsupervised Learning

  • D. d’Humières
Conference paper
Part of the NATO ASI Series book series (volume 20)


Among several models of neural networks(1–4), layered structures are particularly appealing as they lead naturally to a hierarchical representation of the input sets, along with a reduced connectivity between individual cells. In Ref. 3 and 4, it was shown that such layered networks are able to memorize complicated input patterns, such as alphabetic characters, during unsupervised learning. On top of that, the filtering properties of the network can be continuously tuned from very sharp discrimination between similar patterns, to broad class aggregation when the selectivity of the cells is decreased. Unfortunately, it was also shown(4) that these properties are obtained with a reduced stability of the learning (the learning process does not converge for some values of the selectivity).


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Kohonen T. (1978). “Associative Memory” (Springer, New York).Google Scholar
  2. 2.
    Hopfield J.J. (1982). “Neural Networks with emergent collective computational abilities”, Proc. Nat. Ac. Sc. USA, 79, 2554–2558.MathSciNetCrossRefGoogle Scholar
  3. 3.
    Fukushima K. (1975). “Self-Organizing Multilayered Neural Network”, Systems Computers Controls 6, 15–22.Google Scholar
  4. 4.
    d’Humières D. and Huberman B.A. (1984) “Dynamics of Self-Organization in Complex Adaptive Networks”, J. Stat. Phys. 34, 361–379 andzbMATHCrossRefGoogle Scholar
  5. 4a.
    d’Humières D. and Huberman B.A. (1985) in “Dynamical Systems and Cellular Automata” edited by Demongeot J., Golès E. and Tchuente M., Academic Press, 187–195.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1986

Authors and Affiliations

  • D. d’Humières
    • 1
  1. 1.Groupe de Physique de Solides de l’Ecole Normale SupérieureParis Cedex 05France

Personalised recommendations