Measuring Information Flow in Reactive Processes

This paper outlines an approach for measuring information flow within reactive probabilistic systems. First, we present the probabilistic model of reactive labelled transition system with input-output actions. Second, we present the language and semantics for simple reactive processes, and investigate the quantified information flow analysis over this semantics. Third, we define a metric over the semantics and then present a method to compute the leakage in reactive processes. The metric we considered is the square root of the Jensen-Shannon divergence: the quantitative information is contained in the distance between state transformations given by a process metric. Finally, we show that there is a connection between our leakage definition and mutual information in the framework of information theory.