Self-organization properties of a discriminator-based neural network

Organized maps coding sensory information have been found in the brain. T. Kohenen (1988, 1989) showed that it is possible to model a self-organizing system capable of forming a reduced representation of input information into such maps. The work defines a network with identical properties but based on a RAM neuron model, namely the C-discriminator, whereas Kohonen uses the classical linear neuron model. Such RAM models have been studied before, but only within the context of a supervised training scheme, whereas in this work an unsupervised training algorithm is used. The architecture of the network is presented, and the training procedure is defined. It is shown, on a simple example, that the network behavior can be predicted considering the overlap areas between input patterns