Modular Labeling RAAM

  • Alessandro Sperduti
  • Antonina Starita


The Labeling RAAM model is a neural network able to encode labeled graphs in fixed size representations. In order to speed up the training procedure and for reducing in size the developed compressed representations, we propose a modular Labeling RAAM. In order to develop the modular system, we face two main problems: the mapping problem, i.e., how to map components of the structures into modules; and the membership problem, i.e., discovering which module must be used for decoding a compressed representation. The mapping between components and modules can be decided on the basis of the strongly connected components of the structures. The membership problem is solved by resorting to the BAMs derived from each LRAAM module. Preliminary results on the modular system are encouraging.


Hide Layer Abstract Graph Mapping Problem Label Graph Modular System 
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]
    V. Cadoret. Encoding syntactical trees with labelling recursive auto-associative memory. In Proceedings ECAI, 555 (1994). Amsterdam.Google Scholar
  2. [2]
    T. H. Cormen, C. E. Leiserson, and R. L. Rivest. Introduction to Algorithms. The MIT Press, 1990.Google Scholar
  3. [3]
    M. deGerlache, A. Sperduti, and A. Starita. Encoding conceptual graphs by Labeling RAAM. In International Conference on Artificial Neural Networks, 272 (1994). Sorrento.Google Scholar
  4. [4]
    M. deGerlache, A. Sperduti, and A. Starita. Using Labeling RAAM to encode medical conceptual graphs. In International Conference on Neural Networks and Expert Systems in Medicine and Healthcare, 1994. Plymouth.Google Scholar
  5. [5]
    G. E. Hinton. Mapping part-whole hierarchies into connectionist networks. Artificial Intelligence, Vol. 46, 47 (1990).CrossRefGoogle Scholar
  6. [6]
    B. Kosko. Neural Networks and Fuzzy Systems. Prentice Hall, 1992.Google Scholar
  7. [7]
    S. Lonardi, A. Sperduti, and A. Starita. Encoding Pyramids by Labeling RAAM. In IEEE Workshop on Neural Networks for Signal Processing, 651 (1994). Ermioni, Greece.Google Scholar
  8. [8]
    T. Plate. Holographic reduced representations. Technical Report CRG-TR-91-1, Department of Computer Science, University of Toronto, 1991.Google Scholar
  9. [9]
    J. B. Pollack. Recursive distributed representations. Artificial Intelligence, Vol. 46, 77 (1990).CrossRefGoogle Scholar
  10. [10]
    P. Smolensky. Tensor product variable binding and the representation of symbolic structures in connectionist systems. Artificial Intelligence, Vol. 46, 159 (1990).MathSciNetMATHCrossRefGoogle Scholar
  11. [11]
    A. Sperduti. Labeling RAAM. Technical Report 93-029, International Computer Science Institute, 1993. A revised version will appear on Connection Science.Google Scholar
  12. [12]
    A. Sperduti. On some stability properties of the LRAAM model. Technical Report 93-031, International Computer Science Institute, 1993. A revised version will appear on IEEE Transactions on Neural Networks.Google Scholar
  13. [13]
    D. S. Touretzky. Boltzcons: Dynamic symbol structures in a connectionist network. Artificial Intellicence, Vol. 46, 5 (1990).CrossRefGoogle Scholar

Copyright information

© Springer-Verlag/Wien 1995

Authors and Affiliations

  • Alessandro Sperduti
    • 1
  • Antonina Starita
    • 1
  1. 1.Dipartimento di InformaticaUniversità di PisaPisaItaly

Personalised recommendations