Robots As Cognitive Tools Information theoretic analysis of sensory-motor data

Abstract

In this paper we explore the possibility of quantitative analysis of sensory data and the interactions between the signals in different sensory channels as an agent interacts with the real world, in order to get a better intuition of these data which constitute, in a sense, the "raw material" that the neural system has to process. This will give us also a handle on formulating some of the design principles of autonomous agents, which we had worked out in earlier research, in a more quantitative way. As an example of an analysis, we employ information theoretic measures, such as the Shannon Entropy and the Mutual Information. We hope that the insights emerging from this research will eventually lead to a more formal description of some of the design principles of autonomous agents.