Modelling Feature Maps by Attributed Parallel Array Grammars
The notion of n-dimensional attributed parallel array grammars is introduced for describing neural networks. Because of the underlying grid structure, Kohonen’s model of feature maps is especially well suited for being represented by n-dimensional attributed parallel array grammars. Using our formal description model we prove that Kohonen’s global algorithm can be replaced by a local one.
Unable to display preview. Download preview PDF.
- [Fre]FREUND, Rudolf: N-Dimensionale Sprachen. Thesis, Technical University of Vienna, 1982.Google Scholar
- CHKP]HERTZ, John A., KROGH, Anders S., PALMER, Richard G.: Introduction to the Theory of Neural Computation. Santa Fe Institute. Addison Wesley, 1991.Google Scholar
- [Kem]KEMKE, Christel: Modelling Neural Networks by Means of Networks of Finite Automata. Proceedings of the IEEE — First Interntional Conference on Neural Networks, San Diego, CA, 21–24, June (1987).Google Scholar
- [Ko2]KOHONEN, T.: Self-Organization and Associative Memory. Springer Series in Information Sciences 8, Heidelberg, 1984.Google Scholar
- [RMS]RITTER, Helge, MARTINETZ, Thomas, SCHULTEN, Klaus: Neuronale Netze. Addison-Wesley, 1990.Google Scholar