Evolving neural network connectivity

Abstract

The application of evolutionary programming, a stochastic search technique, for determining connectivity in feedforward neural networks is investigated. The method is capable of simultaneously evolving both the connection scheme and the network weights. The number of connections are incorporated into an objective function so that network parameter optimization is done with respect to network complexity as well as mean pattern error. Experimental results are shown for simple binary mapping problems.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>