Vendredi 15 Mars


Retour à la vue des calendrier
Vendredi 15 Mars
Heure: 13:30 - 14:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Linear Dependent Types For Differential Privacy
Description: Marco Gaboardi Differential privacy offers a way to answer queries about
sensitive information while offering strong, provable privacy guarantees.
Several tools have been developed for certifying that a given query is
differentially private. In one approach, Reed and Pierce[31] proposed a
functional programming language, Fuzz, for writing differentially private
queries. Fuzz uses linear types to track sensitivity, as well as a
probability monad to express randomized computation; it guarantees that any
program that has a certain type is differentially private. Fuzz can
successfully verify many useful queries. However, it fails when the
analysis depends on values that are not known statically.
We present DFuzz, an extension of Fuzz with a combination of linear indexed
types and lightweight dependent types. This combination allows a richer
sensitivity analysis that is able to analyze a larger class of queries,
including queries whose sensitivity depends on runtime information. As in
Fuzz, the differential privacy guarantees follows directly from the
soundness theorem for the type system. We demonstrate the enhanced
expressivity of DFuzz by certifying differential privacy a broad class of
iterative algorithms that could not be typed previously. We conclude by
discussing the challenges of DFuzz type checking.