Propagation of Uncertainty in Dependence Graphs

In this paper the main results of DRUMS R.P3.3 Task ’Propagation of Uncertainty in Dependence Graphs’ are presented. It is divided in two parts. In the first one, it is considered a new calculus with imprecise probabilities, which is not as uninformative as classical one, [6, 7, 9]. In the second one, the axiomatic squeme given in [11] has been adapted to Pearl's causal networks, [10], by introducing three new axioms. Then, this axiomatic framework is particularized to the new calculus of upper and lower probabilities. It is shown that upper and lower probabilities may be propagated on causal networks on an analogous way to probabilities.