A new corner classification approach to neural network training

Abstract

The corner classification approach to neural network training has the excellent capability ofprescriptive learning, where the network weights areprescribed merely by inspection of the training samples. This technique is extremely fast compared to other conventional training techniques such as backpropagation. However, the versions described hitherto have been sensitive to the choice of the radius of generalization. We present here a new and improved corner classification technique that retains the prescriptive learning capability and gives excellent generalization performance. This algorithm could be the basis of the recently introduced neuroscientific notion of “working memory.”