How Human Inspired Learning enhances the Behavior of Autonomous Agents

Abstract

An autonomous agent must deal with unforeseen situations that can't be preprogrammed. Therefore, the agent has to make its own experiences, solutions and valuations to situations, actions and objects to be able to enhance previous actions and avoid repeating wrong actions. On the basis of the existing cognitive architecture Simulation of Mental Apparatus and Applications (SiMA) at the Institute of Computer Technology (ICT) an bionically inspired attempt of learning should be implemented in functional model of the human mind which is then used in a multi-agent simulation showing how bionically inspired cognitive architectures can get extended by learning. Due to the attempt in the project SiMA the learning function has to fit in the psychoanalytic model and therefore it has to be compatible with the way of learning that human beings do. This might also help to get a little bit closer to the understanding of how the human mind manipulates memory to show these until now unreached cognitive abilities.