3 papers
cs.LG2020
ReLU activated Multi-Layer Neural Networks trained with Mixed Integer Linear Programs
Steffen Goebbels
In this paper, it is demonstrated through a case study that multilayer feedforward neural networks activated by ReLU functions can in principle be trained iteratively with Mixed In…
math.FA2020
On Sharpness of Error Bounds for Multivariate Neural Network Approximation
Steffen Goebbels
Single hidden layer feedforward neural networks can represent multivariate functions that are sums of ridge functions. These ridge functions are defined via an activation function…
math.FA2018
On Sharpness of Error Bounds for Single Hidden Layer Feedforward Neural Networks
Steffen Goebbels
A new non-linear variant of a quantitative extension of the uniform boundedness principle is used to show sharpness of error bounds for univariate approximation by sums of sigmoid…