An Improved Bound on the VC-Dimension of Neural Networks with Polynomial Activation Functions
Optimization and Control
2007-05-23 v3 Algebraic Geometry
Abstract
In this note, we derive an improved upper bound for the VC-dimension of neural networks with polynomial activation functions. This improved bound is based on a result of Rojas on the number of connected components of a semi-algebraic set.
Keywords
Cite
@article{arxiv.math/0112208,
title = {An Improved Bound on the VC-Dimension of Neural Networks with Polynomial Activation Functions},
author = {J. Maurice Rojas and M. Vidyasagar},
journal= {arXiv preprint arXiv:math/0112208},
year = {2007}
}
Comments
9 pages, submitted for publication. Various typos fixed and the proof of the main result has been streamlined