Wilson, Dennis, Cussat-Blanc, Sylvain and Luga, Hervé (2016) The Evolution of Artificial Neurogenesis. In: Genetic and Evolutionary Computation Conference Companion (GECCO 2016), 20 July 2016 - 24 July 2016, Denver, United States.

Preview

Text Download (1MB) | Preview

Abstract

Evolutionary development as a strategy for the design of artificial neural networks is an enticing idea, with possible inspiration from both biology and existing indirect representations. A growing neural network can not only optimize towards a specific goal, but can also exhibit plasticity and regeneration. Furthermore, a generative system trained in the optimization of the resultant neural network in a reinforcement learning environment has the capability of on-line learning after evolution in any reward-driven environment. In this abstract, we outline the motivation for and design of a generative system for artificial neural network design.