Robustness

Abstract

Given a channel with input X, we have now seen that g-leakage provides us with a rich variety of ways to measure the information leakage of X that the channel causes. The prior models the adversary’s prior knowledge about X; the gain function g models the operational scenario, which encompasses both the set of actions that the adversary can take and also the worth to the adversary of each such action, for each possible value of X; and the choice of multiplicative- or additive leakage allows us to measure either the relative or the absolute increase in g-vulnerability.