On the approximation capability of shallow and deep neural networks having smooth activations with respect to the Sobolev norm.

Park, Hyeokjoo · Neural Netw · 2026

basic_science · Level V

Where this comes from

Abstract

In this paper, we investigate the simultaneous approximation of the functions and their derivatives by neural networks having smooth non-polynomial activation functions and a fixed depth, which is motivated by the physics-informed machine learning. We start by proving that the neural networks with smooth non-polynomial activation functions and with only one hidden layer having width O(N<sup>d</sup>) can approximate any W<sup>s,p</sup>-regular function with rate O(N<sup>k-s</sup>) in the W<sup>k,p</sup>-norm. We then extend this result to the networks having more than one hidden layers by using the mathematical induction.