Bell and experimental tests of his inequality showed that it is impossible to explain all of the predictions of quantum mechanics using a theory which satisfies the basic concepts of locality and realism, but which (if not both) is violated is still an open question. As it seems impossible to resolve this question experimentally, one can ask how plausible realism -- the idea that external properties of systems exist prior to and independent of observations -- is, by considering the amount of resources consumed by itself and its non-local features. I will construct an explicit realistic model in which the number of hidden-variable states scales polynomially with the number of possible quantum measurements. In the limit of a large number of measurements, the model recovers the result of Montina, that no hidden-variable theory that agrees with quantum predictions could use less hidden-variable states than the straightforward model in which every quantum state is associated with one such hidden state. Thus, for any given system size, realistic theories cannot describe nature more efficiently than quantum theory itself. I will then turn to the problem of "non-locality" in realistic theories showing that every such theory that agrees with quantum predictions allows superluminal signaling at the level of hidden variable states.


Talk Number PIRSA:09090084
Speaker Profile Časlav Brukner