Abstract

It is well known that the addition of noise to the input data of a neural network during training can, in some circumstances, lead to significant improvements in generalization performance. Previous work has shown that such training with noise is equivalent to a form of regularization in which an extra term is added to the error function. However, the regularization term, which involves second derivatives of the error function, is not bounded below, and so can lead to difficulties if used directly in a learning algorithm based on error minimization. In this paper we show that, for the purposes of network training, the regularization term can be reduced to a positive definite form which involves only first derivatives of the network mapping. For a sum-of-squares error function, the regularization term belongs to the class of generalized Tikhonov regularizers. Direct minimization of the regularized error function provides a practical alternative to training with noise. 1 Regularization...

Citations

44	Creating artificial neural networks that generalize - Sietsma, Dow - 1991
40	Exact calculation of the hessian matrix for the multilayer perceptron - Bishop - 1992
39	Improving the Generalization Properties of Radial Basis Function Neural Networks - Bishop - 1991
2	Curvature-Driven Smoothing in Feed-forward Networks - Bishop - 1993
2	Noise injection into inputs in back-propagation training - Matsuoka - 1992