Beyond Differential Privacy: Composition Theorems and Relational Logic for f-divergences between Probabilistic Programs

By Gilles Barthe, Federico Olmedo. 2013

99 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Privacy & Security, Information Theory

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Privacy & Security

In Information Theory

Citations in the corpus

Detecting Violations of Differential Privacy (2018, background): “On the other hand, static methods exist for verifying that a program obeys differential privacy during any execution, based on relational program logic [1–7] and relational type system [24, 35, 41].”
Advanced Probabilistic Couplings for Differential Privacy (2016, positive): “…of the differential privacy guarantee, using a wide variety of veriﬁcation approaches: dynamic checking [24, 36], relational program logic [2, 6] and relational reﬁnement type systems [9], linear (dependent) type systems [29, 42], product programs [7], meth-ods based on computing…”
Proving Differential Privacy via Probabilistic Couplings (2016, neutral): “There are seemingly few applications of coupling in formal verification, despite considerable research on probabilistic bisimulation (first introduced by Larsen and Skou (1989)) and probabilistic relational program logics (first introduced by Barthe et al. (2009)).”
Privacy Amplification by Subsampling: Tight Analyses via Couplings and Divergences (2018, background): “Theorem 1 ([Barthe and Olmedo, 2013 Note that in the statement of the theorem we take α = e ε .”
A Better Bound Gives a Hundred Rounds: Enhanced Privacy Guarantees via f-Divergences (2020, unknown): “It is shown in [27], [28] that M is (ε, δ)-DP⇐⇒ sup d∼d′ Eeε(Md‖Md′) ≤ δ.”
Probabilistic Relational Reasoning for Differential Privacy (2013, background): “Moreover , we say that a program c is lossless iff c m = 1 for any initial memory m.”
Obfuscation via Information Density Estimation (2019, background): “g, [49, 50]), in statistics as the probability of correct decision in Bayesian binary hypothesis testing [48], and in information theory for proving general channel coding converse results [48,51].”
Differentially Private Federated Learning: An Information-Theoretic Perspective (2021, methodology): “…Neyman-Pearson lemma we can obtain an alternative formula for implying that the DP constraint (1) can be equivalently expressed in terms of E ε -divergence [26]: This E ε -divergence representation of DP was used in [27–31] to prove new privacy results or simplify the proofs of existing results.”
Hypothesis Testing Interpretations and Renyi Differential Privacy (2019, methodology): “We can extend this result to general measurable setting by using the continuity of ∆ ε (see also [Liese and Vajda, 2006]), which is obtained by f -divergence characterization of ∆ ε [Barthe and Olmedo, 2013].”