Growing functional modules, a prospective paradigm for epigenetic artificial intelligence

Abstract

Epigenesis postulates that intelligence may arise in a system thanks to an adequate developmental process interacting with its environment. Robotics constitutes the most suitable field to implement this approach. This paper presents the underlying concept, an earlier proposal and a potential implementation of a prospective paradigm named Growing Functional Modules. This paradigm has been conceived to allow the design and the automatic development of a distributed and dynamic architecture able to gradually control an epigenetic robot. An illustration is given applying the paradigm to create a simple artificial brain for a mushroom shaped robot. © Springer-Verlag Berlin Heidelberg 2005.