Seminar: Dr. Nihat Ay, Information Maximization in Recurrent Structures (12/9/2008)

Dr. Nihat Ay,  Max Planck Institute for Mathematics in the Sciences, Leipzig, Germany.

Title: Information Maximization in Recurrent Structures

Abstract: In the first part of my talk, I present results on the maximization of a complexity measure that is closely related to information flow maximization in recurrent networks. These results refer to the emergence of particular spatio-temporal patterns. In order to study the behavioural implications of such an optimization principle one has to embed the network in a perception-action loop. I address this issue in the second part of my talk and demonstrate the emergence of explorative and coordinated behaviour as consequnce of predictive information maximization.