Generalised Context Privacy

Talk at Manchester University, Manchester, UK
Project: Generalised Context Privacy

There has been significant effort on addressing data content threats, whether this be to data in motion, data at rest or a private release of data. However, for systems being directly observed by an adversary, risks exist in the form of information revealed by the actions a system takes and the context in which it takes them. Thus, considering the context privacy of a system is important from a holistic viewpoint. Context privacy techniques already exist in a wide range of different domains (e.g., onion routing, location privacy in ad-hoc networks), but we lack the ability to rapidly develop novel techniques when new context privacy threats are identified. To address this, privacy quantification approaches and techniques will be developed for an arbitrary system, which will subsequently be used in translators for specific systems of interest. This presentation sets out the problem and explores initial work on quantification being performed in a recently funded EPSRC project.