Design of Modular Neural Network Architectures Using Genetic Algorithms

In this paper, we propose an evolutionary approach to the design of optimal modular neural network architectures. In this approach, a modular neural network is treated as a phenotype of an individual, and the modular architecture is optimized through the evolution of its genetic representation (genotype) by using genetic algorithms. As one of the modular neural networks, we adopt Cross-Coupled Hopfield Nets (CCHN) in which plural Hopfield networks are coupled to each other. The architecture of the CCHN is represented by some structural-parameters such as the number of modules, the numbers of module units, and the module connectivity. These parameters for an individual are encoded in a binary string. In the simulation, our genetic system is applied to associative memories. The fitness of an individual is defined so as to be larger when the individual has a simpler architecture as well as when the association performance is higher. In the simulation, we verify that the genetic system finds highperformance individuals with simple modular architectures.