Probabilistic Inference and Monadic Second Order Logic

International audienceThis paper combines two classic results from two different fields: the result by Lauritzen and Spiegelhalter [21] that the probabilistic inference problem on probabilistic networks can be solved in linear time on networks with a moralization of bounded treewidth, and the result by Courcelle [10] that problems that can be formulated in counting monadic second order logic can be solved in linear time on graphs of bounded treewidth. It is shown that, given a probabilistic network whose moralization has bounded treewidth and a property P of the network and the values of the variables that can be formulated in counting monadic second order logic, one can determine in linear time the probability that P holds