A backpropagation and initialization routine for hyperbolic sigma-pi neural networks

Authors:
Burkhard Lenze
Affiliations:
Department of Computer Science, University of Applied Sciences, Fachhochschule Dortmund, 44047 Dortmund, Germany
Venue:
Neural, Parallel & Scientific Computations
Year:
2003

Citing 11
Cited 0

Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations

Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations
TDI-subspaces of CRd and some density problems from neural networks

Journal of Approximation Theory
Adaptive critic for sigma-pi networks

Neural Networks
Partially pre-calculated weights for the backpropagation learning regime and high accuracy function mapping using continuous input RAM-based sigma-pi nets

Neural Networks
Transformations of sigma-pi nets: obtaining reflected functions by reflecting weight matrices

Neural Networks
An Adaptive Activation Function for Higher Order Neural Networks

AI '02 Proceedings of the 15th Australian Joint Conference on Artificial Intelligence: Advances in Artificial Intelligence
Long-term attraction in higher order neural networks

IEEE Transactions on Neural Networks
Lp approximation of Sigma-Pi neural networks

IEEE Transactions on Neural Networks
Learning capability assessment and feature space optimization for higher-order neural networks

IEEE Transactions on Neural Networks
Generalization in multi-layer networks of Sigma-pi units

IEEE Transactions on Neural Networks
A framework for improved training of Sigma-Pi networks

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

As it is well-known, sigma-pi neural networks offer an interesting alternative to classical ridge-type neural networks, especially when aspects of invariances enter the field. In this contribution, we develop a backpropagation-type algorithm for a special kind of these networks, so-called hyperbolic sigma-pi neural networks, together with a well-founded initialization routine for finding proper starting parameters. We will not only sketch the mathematical machinery but also give detailed pseudo code for rapid implementation.