You are warmly invited to attend our upcoming statistics seminar hosted by the Department of Mathematics.

On 18 June 2026, Professor Christian Robert from Université Paris Dauphine will give a talk on 'Bayesian Adversarial Privacy.'

This talk will take place in person at King's College London in Room G40 in the King's Building from 14:15.

This event is open to all, but to ensure smooth entry on the day, we ask that attendees who are not current members of staff at King's to please click 'Register for this event' to sign up for an In-Person Ticket. 

 

Abstract

Bayesian Adversarial Privacy

Theoretical and applied research into privacy encompasses an incredibly broad swathe of differing approaches, emphases, and aims. In the first part, we propose a novel framework for measuring privacy from a Bayesian game-theoretic perspective. This framework enables the creation of new, purpose-driven privacy definitions that are rigorously justified, while also allowing for the assessment of existing privacy guarantees through game theory. We show that pure and probabilistic differential privacy are special cases of our framework and provide new interpretations of the post-processing inequality in this setting. Further, we demonstrate that privacy guarantees can be established for deterministic algorithms, which are overlooked by current privacy standards.

In the second part, we introduce a new quantitative notion of privacy that is both contextual and specific. We argue that it provides a more meaningful notion of privacy than the widely utilised framework of differential privacy and a more explicit and rigorous formulation than what is commonly used in statistical disclosure theory. Our definition relies on concepts inherent to standard Bayesian decision theory, while departing from it in several important respects. In particular, the party controlling the release of sensitive information should make disclosure decisions from the prior viewpoint, rather than conditional on the data, even when the data is itself observed.

For further details on the joint work with James Bailie, Cameron Bell, Joshua Bon, Timothy Johnston, Antoine Luciano, and Judith Rousseau, see the recent preprints on Persuasive Privacy and Bayesian Adversarial Privacy.

 

About the Speaker: 

Christian Robert is currently a Professor at Université Paris-Dauphine (France) and a part-time Professor at the University of Warwick, Department of Statistics. He is a senior member of the Institut Universitaire de France (IUF) and a member of the Statistics Laboratory at Centre de Recherche en Economie et Statistique (CREST) in Paris-Saclay. His research focuses on Bayesian statistics, computational statistics and latent variable method. His interests include decision theory, model choice, foundations, objective Bayesian methodology, Monte Carlo methodology, MCMC methods, sequential importance sampling, and approximate Bayesian computation. He is a former Editor of the JRSSB and a current Deputy-Editor for Biometrika. He was President of the International Society for Bayesian Analysis (ISBA) in 2008 and is a Fellow of the IMS, the ISBA and the ASA.