Using Genetic Engineering to Find Modular Structures and Activation Functions for Architectures of Artifical Neural Networks

摘要

An Evolutionary Algorithm is used to optimize the archtiecture and activation functions of an Artificial Neurla Networks. IT wil be shown that it is possible, with the help of a graph-database and Genetic Engineering, to find modular structures for these networks. Some new graph-rewritings are used to construct families of architectures from these modular structures. Simulation results for two problems are given. An analysis of the data in the databae suggest the usage of symmetric activation functions.