Robust and private Bayesian inference. (English) Zbl 1432.68132
Auer, Peter (ed.) et al., Algorithmic learning theory. 25th international conference, ALT 2014, Bled, Slovenia, October 8–10, 2014. Proceedings. Berlin: Springer. Lect. Notes Comput. Sci. 8776, 291-305 (2014).
Summary: We examine the robustness and privacy of Bayesian inference, under assumptions on the prior, and with no modifications to the Bayesian framework. First, we generalise the concept of differential privacy to arbitrary dataset distances, outcome spaces and distribution families. We then prove bounds on the robustness of the posterior, introduce a posterior sampling mechanism, show that it is differentially private and provide finite sample bounds for distinguishability-based privacy under a strong adversarial model. Finally, we give examples satisfying our assumptions.
For the entire collection see [Zbl 1297.68011].
For the entire collection see [Zbl 1297.68011].
MSC:
68P27 | Privacy of data |
62F15 | Bayesian inference |
68T05 | Learning and adaptive systems in artificial intelligence |