Pattern Classification in a Perceptron Circuit Using Lead Phthalocyanine Films

Evaporated films of lead phthalocyanine (PbPc) show a unique conductivity change after a very slight amount of gases are adsorbed. The change in the conductivity of the films depends on the polarity and the strength of input voltages, and therefore is analogous to the change of synaptic connections in nerve systems. Therefore, a perceptron circuit has been fabricated using PbPc films as synaptic connections. After learning via loading of input signals and teacher signals, this circuit can classify random inputs of eight-dimensional voltage patterns into two categories when the input patterns are linearly separable.