Approximation in Lp(μ) with deep ReLU neural networks

We discuss the expressive power of neural networks which use the non-smooth ReLU activation function varrho(x) = max0,x by analyzing the approximation theoretic properties of such networks. The existing results mainly fall into two categories: approximation using ReLU networks with a fixed depth, or using ReLU networks whose depth increases with the approximation accuracy. After reviewing these findings, we show that the results concerning networks with fixed depth--- which up to now only consider approximation in Lp(λ) for the Lebesgue measure λ--- can be generalized to approximation in Lp(μ), for any finite Borel measure μ. In particular, the generalized results apply in the usual setting of statistical learning theory, where one is interested in approximation in L2(mathbbP), with the probability measure mathbbP describing the distribution of the data.