Aktive Computation in offenen Systemen. Lerndynamiken in biologischen Systemen: Vom Netzwerk zum Organismus

Abstract

In this thesis the action-perception-cycle, which is inherent in every cognitive system, is investigated by several special examples. These examples are taken from different levels of description and are biological motivated. Moreover, a novel Hebb-like learning rule for neural networks is introduced, which has not only interesting features by itself but establishes a connection to the action-perception-cycle and demonstrates the interplay between different levels of description.