Abstract
We show that neural networks with an absolute value activation function and with network path norm, network sizes and network weights having logarithmic dependence on 1/ε can ε-approximate functions that are analytic on certain regions of Cd.
Subject
General Physics and Astronomy
Reference21 articles.
1. Universal Approximation Using Feedforward Neural Networks: A Survey of Some Existing Methods, and Some New Results
2. The expressive power of neural networks: A view from the width;Lu;Adv. Neural Inf. Process. Syst.,2017
3. Exponential convergence of the deep neural network approximation for analytic functions
4. Norm-based capacity control in neural networks;Neyshabur;Proceedings of the 28th Conference on Learning Theory (COLT),2015
5. Nonparametric regression using deep neural networks with ReLU activation function;Schmidt-Hieber;Ann. Stat.,2020
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献