Self-organization of associative memory and pattern classification: recurrent signal processing on topological feature maps
- 234 Downloads
We extend the neural concepts of topological feature maps towards self-organization of auto-associative memory and hierarchical pattern classification. As is well-known, topological maps for statistical data sets store information on the associated probability densities. To extract that information we introduce a recurrent dynamics of signal processing. We show that the dynamics converts a topological map into an auto-associative memory for real-valued feature vectors which is capable to perform a cluster analysis. The neural network scheme thus developed represents a generalization of non-linear matrix-type associative memories. The results naturally lead to the concept of a feature atlas and an associated scheme of self-organized, hierarchical pattern classification.
KeywordsNeural Network Cluster Analysis Signal Processing Probability Density Feature Vector
Unable to display preview. Download preview PDF.
- Kohonen T (1984) Self-organization and associative memory. Springer, Berlin Heidelberg New YorkGoogle Scholar
- Kühnel H (1990) Diplomarbeit, Physik-Department, Technische Universität MünchenGoogle Scholar
- Ritter H (1989) Asymptotic level density for a class of vector quantization processes. Technical Report, University of HelsinkiGoogle Scholar
- Ritter H, Schulten K (1988a) Extending Kohonen's self-organizing mapping algorithm to learn ballistic movements. In: Eckmiller R, Malsburg C von der (eds) Neural computers. Springer, Berlin Heidelberg New York, pp 393–406Google Scholar
- Ritter H, Schulten K (1988b) Kohonen's self-organizing maps: exploring their computational capabilities. IEEE ICNN 88 Conference, San Diego, pp 109–116Google Scholar
- Willshaw DJ, Malsburg C von der, (1976) How patterned neural connections can be set up by self-organization. Proc R Soc London B 194:431–445Google Scholar