Associative Storage and Retrieval of Digital Information in Networks of Adaptive “Neurons”

An adaptive logical element, called the ADALINE “neuron,” which consists of a set of variable weights, a threshold, and adaptation machinery for automatically adjusting its weights, has been described previously.* Proofs of convergence of its learning processes and derivations of learning rates have been made. It has been demonstrated analytically and empirically that a single ADALINE can be “trained” to recognize geometric patterns, perform logical functions, and store digital information.