Engineering cognitive systems- some conceptual issues

The purpose of this paper is to stimulate serious reflection and discussion on the problems associated with the design and implementation of complex adaptable systems and provide some conceptual tools for taking a number of research issues further. Some general comments will be made about symbolic and connectionist system with the focus to some degree on the latter category. We shall take connectionist systems as those in which the interaction of many simple information processing components bring about the effect of learning and adaptation. This includes a broad class of computational architectures and this paper will reference artificial neural networks, classifier systems, evolutionary algorithms and immune algorithms