Approximation Analysis of Convolutional Neural Networks

Abstract

In its simplest form, convolution neural networks (CNNs) consist of a fully connected two-layer network g composed with a sequence of convolution layers T. Although g is known to have the universal approximation property, it is not known if CNNs, which have the form g ◦ T inherit this property, especially when the kernel size in T is small. In this paper, we show that under suitable conditions, CNNs do inherit the universal approximation property and its sample complexity can be characterized. In addition, we discuss concretely how the nonlinearity of T can improve the approximation power. Finally, we show that when the target function class has a certain compositional form, convolutional networks are far more advantageous compared with fully connected networks, in terms of the number of parameters needed to achieve the desired accuracy.