On the Additive Capacity Problem for Quantitative Information Flow

International audiencePreventing information leakage is a fundamental goal in achieving confidentiality. In many practical scenarios, however, eliminating such leaks is impossible. It becomes then desirable to quantify the severity of such leaks and establish bounds on the threat they impose. Aiming at developing measures that are robust wrt a variety of operational conditions, a theory of channel capacity for the g-leakage model was developed in [1], providing solutions for several scenarios in both the multiplicative and the additive setting. This paper continuous this line of work by providing substantial improvements over the results of [1] for additive leakage. The main idea of employing the Kan-torovich distance remains, but it is now applied to quasimetrics, and in particular the novel " convex-separation " quasimetric. The benefits are threefold: first, it allows to maximize leakage over a larger class of gain functions, most notably including the one of Shannon. Second, a solution is obtained to the problem of maximizing leakage over both priors and gain functions, left open in [1]. Third, it allows to establish an additive variant of the " Miracle " theorem from [3]