On the existence of optimal shallow feedforward networks with ReLU activation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dereich, Steffen, Kassing, Sebastian
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916487355170816
author Dereich, Steffen
Kassing, Sebastian
author_facet Dereich, Steffen
Kassing, Sebastian
contents We prove existence of global minima in the loss landscape for the approximation of continuous target functions using shallow feedforward artificial neural networks with ReLU activation. This property is one of the fundamental artifacts separating ReLU from other commonly used activation functions. We propose a kind of closure of the search space so that in the extended space minimizers exist. In a second step, we show under mild assumptions that the newly added functions in the extension perform worse than appropriate representable ReLU networks. This then implies that the optimal response in the extended target space is indeed the response of a ReLU network.
format Preprint
id arxiv_https___arxiv_org_abs_2303_03950
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle On the existence of optimal shallow feedforward networks with ReLU activation
Dereich, Steffen
Kassing, Sebastian
Machine Learning
Numerical Analysis
Optimization and Control
Primary 68T07, Secondary 68T05, 41A50
We prove existence of global minima in the loss landscape for the approximation of continuous target functions using shallow feedforward artificial neural networks with ReLU activation. This property is one of the fundamental artifacts separating ReLU from other commonly used activation functions. We propose a kind of closure of the search space so that in the extended space minimizers exist. In a second step, we show under mild assumptions that the newly added functions in the extension perform worse than appropriate representable ReLU networks. This then implies that the optimal response in the extended target space is indeed the response of a ReLU network.
title On the existence of optimal shallow feedforward networks with ReLU activation
topic Machine Learning
Numerical Analysis
Optimization and Control
Primary 68T07, Secondary 68T05, 41A50
url https://arxiv.org/abs/2303.03950