On model selection and the disability of neural networks to decompose tasks
Adaptation and Self-Organizing Systems
2007-05-23 v1 Disordered Systems and Neural Networks
Neural and Evolutionary Computing
Abstract
A neural network with fixed topology can be regarded as a parametrization of functions, which decides on the correlations between functional variations when parameters are adapted. We propose an analysis, based on a differential geometry point of view, that allows to calculate these correlations. In practise, this describes how one response is unlearned while another is trained. Concerning conventional feed-forward neural networks we find that they generically introduce strong correlations, are predisposed to forgetting, and inappropriate for task decomposition. Perspectives to solve these problems are discussed.
Keywords
Cite
@article{arxiv.nlin/0202038,
title = {On model selection and the disability of neural networks to decompose tasks},
author = {Marc Toussaint},
journal= {arXiv preprint arXiv:nlin/0202038},
year = {2007}
}
Comments
LaTeX, 7 pages, 3 figures