Fig. 1From: Superiority of quadratic over conventional neural networks for classification of gaussian mixture dataNeural network training as a computational optimization process with respect to an objective function which is the error rate for classification, without loss of generalityBack to article page