Flexibility of generalization studies with a neuromolecular model: towards learning generalization

An abstract neuromolecular neuron model (called the cytomatrix neuron) that illustrates the structure-function plasticity similar to the one in biological organizations is described. The cytomatrix neuron is a softened cellular automaton with molecular components, roughly motivated by interactions that could occur in a molecular or cellular complex. A multiparameter evolutionary algorithm that acts on the various parameters is used for learning. Experiments with various type bit-pattern learning tasks demonstrate that the flexibility of generalization can be increased. This increase in variability of responses can be exploited further. A multiparameter evolutionary algorithm for learning generalization is presented and discussed.