INB Lunch-Seminar

On the Concept of Information Flow

Dr. Nihat Ay
Leiter der Arbeitsgruppe "Information Theory of Cognitive Systems"
am Max-Planck-Institut für Mathematik in den Naturwissenschaften,
Leipzig

Information theory provides a useful framework for the formulation of firstlearning principles in neural networks. In feed-forward structures, the interpretation of the corresponding objective functions is usually closely related to the causal concept of information flow. Within the context of more general recurrent networks, information-theoretic measures quantify the stochastic dependence among parts of the system and allow for a study of the generated phenomenological system complexity. In contrast to the feed-forward situation this is not directly linked to information flows. In my talk, I want to present some results on the relation between stochastic dependence, information flows, and complexity.

Zeit: Freitag, den 29. Juni 2007, 12 Uhr c.t.

Ort: Institut für Neuro- und Bioinformatik
        Seminarraum (1. OG, Raum 17),
        Ratzeburger Allee 160 (Geb. 64, 1. OG)