Abstract
We investigate the convergence speed of the Self Organizing Map (SOM) and Dynamic Link Matching (DLM) on a benchmark problem for the solution of which both algorithms are good candidates. We show that the SOM needs a large number of simple update steps and DLM a small number of complicated ones. A comparison of the actual number of floating point operations hints at an exponential vs. polynomial scaling behavior with increased pattern size. DLM turned out to be much less sensitive to parameter changes than the SOM.
Funding from the HCM network “Parallel modeling of neural operators for pattern recognition” by the European Community is gratefully acknowledged.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
R.P. Würtz. Multilayer Dynamic Link Networks for Establishing Image Point Correspondences and Visual Object Recognition, volume 41 of Reihe Physik. Verlag Harri Deutsch, Thun, Frankfurt am Main, 1995.
W. Konen and C.v.d. Malsburg. Learning to generalize from single examples in the dynamic link architecture. Neural Computation, 5:719–735, 1993.
T.J. Sejnowski, P.K. Kienker, and G.E. Hinton. Learning symmetry groups with hidden units: Beyond the perceptron. Physica D, 22:260–275, 1986.
D.J. Willshaw and C. v.d. Malsburg. How patterned neural connections can be set up by self-organization. Proceedings of the Royal Society, London B, 194:431–445, 1976.
W. Konen, T. Maurer, and C. v.d. Malsburg. A fast dynamic link matching algorithm for invariant pattern recognition. Neural Networks, 7(6/7):1019–1030, 1994.
T. Kohonen. Self-organized formation of topologically correct feature maps. Biological Cybernetics, 43:59–69, 1982.
T. Kohonen. The self-organizing map. Proc. IEEE, 78:1464–1480, 1990.
K.-O. Behrmann. Leistungsuntersuchungen des Dynamischen-Link-Matchings und Vergleich mit dem Kohonen-Algorithmus. Technical Report IR-INI 93-05, Ruhr-Universität Bochum, 1993.
R.P. Würtz, W. Konen, and K.-O. Behrmann. On the performance of neuronal matching algorithms. Manuscript in preparation.
Author information
Authors and Affiliations
Corresponding author
Editor information
Rights and permissions
Copyright information
© 1996 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Würtz, R.P., Konen, W., Behrmann, KO. (1996). How fast can neuronal algorithms match patterns?. In: von der Malsburg, C., von Seelen, W., Vorbrüggen, J.C., Sendhoff, B. (eds) Artificial Neural Networks — ICANN 96. ICANN 1996. Lecture Notes in Computer Science, vol 1112. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-61510-5_28
Download citation
DOI: https://doi.org/10.1007/3-540-61510-5_28
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-61510-1
Online ISBN: 978-3-540-68684-2
eBook Packages: Springer Book Archive