Generalised probabilistic theories meet Bayesian networks

Wednesday, May 14, 2014, 4:15pm – 5:00pm
Presented by Raymond Lal
Located in LAK 2.06

Bayesian networks provide a powerful tool for reasoning about probabilistic causation, used in many areas of science.  They are, however, intrinsically classical.  In particular, Bayesian networks naturally yield the Bell inequalities.  Inspired by this connection, we generalise the formalism of classical Bayesian networks in order to investigate non-classical correlations in arbitrary causal structures.  Our framework of `generalised Bayesian networks’ replaces latent variables with the resources of any generalised probabilistic theory, most importantly quantum theory, but also, for example, Popescu-Rohrlich boxes.

We obtain three main sets of results.

Firstly, we prove that all of the observable conditional independences required by the classical theory also hold in our generalisation; to obtain this, we extend the classical $d$-separation theorem to our settting. Secondly, we find that the theory-independent constraints on probabilities can go beyond these conditional independences.  For example we find that no probabilistic theory predicts perfect correlation between three parties using only bipartite common causes.  Finally, we begin a classification of those causal structures, such as the Bell scenario, that may yield a separation between classical, quantum and general-probabilistic correlations.

Joint work with Joe Henson and Matt Pusey, based on the following paper: