A learning rule for very simple universal approximators consisting of a single layer of perceptrons.

A learning rule for very simple universal approximators consisting of a single layer of perceptrons.