A Note on Differential Privacy: Defining Resistance to Arbitrary Side Information

By Shiva Prasad Kasiviswanathan, Adam D. Smith. 2008

103 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

The structure of optimal private tests for simple hypotheses (2018, methodology): “We study algorithms that satisfy differential privacy (DP) [DMNS06], a restriction on the al-gorithm that ensures meaningful privacy guarantees against an adversary with arbitrary side information [KS08].”
Differentially Private Empirical Risk Minimization: Efficient Algorithms and Tight Error Bounds (2014, background): “See Dwork [18], Kasiviswanathan and Smith [30], Kifer and Machanavajjhala [33] for discussion of the “semantics” of differential privacy.”
Universally utility-maximizing privacy mechanisms (2008, background): “Below we discuss the papers in the diﬀerential privacy literature closest to the present work; see [9] for a recent, thorough survey of the state of the ﬁeld.”
Composition attacks and auxiliary information in data privacy (2008, methodology): “In this paper we just reproduce the relevant parts from [22].”
Differentially Private Confidence Intervals (2020, background): “Interested readers should see [11] or [2] for more discussion.)”
Pufferfish (2014, methodology): “P uffer F ish( S , pairs ) satisﬁes the axioms of convexity and transformation invariance.”
Efficient, Differentially Private Point Estimators (2008, background): “meaning of differential privacy—see Section 4, Remark 2 in [16] and [ 25 ] for discussion.”
Differentially private data aggregation with optimal utility (2014, background): “In this work we focus on the notion of differential privacy, but for completeness, we refer to some recent papers that investigate limitations of this notion [37,43] and propose alternative definitions of privacy for statistical queries [15,33,42].”
Differentially Private Filtering (2012, background): “In contrast, differential privacy is a worst-case notion that holds independently of any probabilistic assumption on the dataset, and provides guarantees (different from those of [17], [18]) against adversaries with arbitrary side information [9].”
Differentially Private Matrix Completion, Revisited (2017, background): “There are semantic reasons for such choice of parameters [29], but that is beyond the scope of the current work.”
Coupled-Worlds Privacy: Exploiting Adversarial Uncertainty in Statistical Data Privacy (2013, background): “” This natural-language interpretation was formalized by [9], [14] in terms of a Bayesian attacker who starts with some prior distribution on the dataset and, based on the output of the mechanism, draws inferences about xi.”