A developmental model of neural computation using cartesian genetic programming

Abstract

The brain has long been seen as a powerful analogy from which novel computational techniques could be devised. However, most artificial neural network approaches have ignored the genetic basis of neural functions. In this paper we describe a radically different approach. We have devised a compartmental model of a neuron as a collection of seven chromosomes encoding distinct computational functions representing aspects of real neurons. This model allows neurons, dendrites, and axon branches to grow, die and change while solving a computational problem. This also causes the synaptic morphology to change and affect the information processing. Since the appropriate computational equivalent functions of neural computation are unknown, we have used a form of genetic programming known as Cartesian Genetic Programming (CGP) to obtain these functions. We have evaluated the learning potential of this system in the context of solving a well known agent based learning scenario, known as wumpus world and obtained promising results.

Study of estimation of distribution algorithms applied to neuroevolution

“…In this study, we used the solution representations of GEP and CGP because they are easily mapped to traditional EDA, and they have a many-to-one genotype to phenotype mapping. As shown in [4,17,19], linear representations that support a many-to-one mapping between the structures subject to genetic modification and the structures subject to selection have become instrumental for the evolution of highly complex structures, in particular ANN.…”

“…In CGP-ANN, the authors of [19] used a 17-bit representation for the weights. However, they did not specify how this representation worked, so we have provided our own method.…”

Study of estimation of distribution algorithms applied to neuroevolution

“…In this study, we used the solution representations of GEP and CGP because they are easily mapped to traditional EDA, and they have a many-to-one genotype to phenotype mapping. As shown in [4,17,19], linear representations that support a many-to-one mapping between the structures subject to genetic modification and the structures subject to selection have become instrumental for the evolution of highly complex structures, in particular ANN.…”

“…In CGP-ANN, the authors of [19] used a 17-bit representation for the weights. However, they did not specify how this representation worked, so we have provided our own method.…”