(1) Max Planck Institute for Mathematics in the Sciences, Inselstr. 22–26, 04103 Leipzig, Germany;(2) Santa Fe Institute, 1399 Hyde Park Road, 87501 Santa Fe, NM, USA
Abstract:
We introduce concepts of external and internal complexity to analyze the relation between an adaptive system and its environment.
We apply this theoretical framework to the construction of models in a cognitive system and the selection between hypotheses
through selective observations performed on a data set in a recurrent process and propose a corresponding neural network architecture.