Dear all, Note an update of today's talk location. Speaker: Manuel Forster Title: Towards identifying a universal unit of nonlocality Time: 5 pm Place: IFW D 42 Abstract: As a universal unit of nonlocality we understand a system allowing the construction of non-interactive protocols to simulate any nonsignaling system. Could such a unit be bipartite? The answer is no - for example no bipartite system can simulate correlations arising from measurements on a five-qubit cluster state. I will explain Barrett and Pironio's nice and simple proof of this statement. Best regards. Manuel