Differential privacy and robust statistics

By Cynthia Dwork, Jing Lei. 2009

830 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Science Applications, Statistics and Probability, Privacy & Security

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 Computer Science Applications

In Statistics and Probability

In Privacy & Security

Citations in the corpus

Certified Robustness to Adversarial Examples with Differential Privacy (2018, background): “It is shown that learning algorithms that satisfy DP with respect to the training data have statistical benefits in terms of out-of-sample performance ; or that DP has a deep connection to robustness at the dataset level [14], [16].”
AdaCliP: Adaptive Clipping for Private SGD (2019, background): “For the Gaussian mechanism, [14] introduced moments accountant, which provides tighter privacy bounds compared to other composition theorems [33, 34, 35, 34, 36].”
Discrete Distribution Estimation under Local Privacy (2016, background): “…also appears in the global privacy context where a trusted service provider releases randomized data (e.g., NIH releasing medical records) to protect sensitive user information (Dwork, 2006; Dwork et al., 2006; Dwork & Lei, 2009; Dwork, 2008; Diakoniko-las et al., 2015; Blocki et al., 2016).”
Extremal Mechanisms for Local Differential Privacy (2014, background): “Differential privacy has been proposed in the global privacy context to formally capture this notion of privacy [11, 13, 12].”
The Composition Theorem for Differential Privacy (2013, methodology): “A routine argument shows that the composition of k queries, each of which is (ε, δ)differentially private, is at least (k , kδ)-differentially private (Dwork et al., 2006b;a; Dwork & Lei, 2009; Dwork et al., 2010). A tighter bound of (ε̃δ̃, kδ + δ̃)-differential privacy under k-fold adaptive composi”
Privacy-Preserving Ridge Regression on Hundreds of Millions of Records (2013, neutral): “, [55], [56]) within the regression circuit.”
Boosting and Differential Privacy (2010, neutral): “[10] for ε-differential privacy and [9], [19], [8] for (ε, δ)-differential privacy).”
Advanced Probabilistic Couplings for Differential Privacy (2016, neutral): “We demonstrate these rules on a classic example of accuracy-dependent privacy: the Propose-test-release framework [16, 48].”
To Drop or Not to Drop: Robustness, Consistency and Differential Privacy Properties of Dropout (2015, methodology): “By the deﬁnition of distance of a dataset to instability , and the composition property Dwork and Lei [2009] of differential privacy we know that for any data set D (cid:48) differing in at most m Γ records from D , there exists a set Q s.t. both over the randomness of A ( D ) and over the…”
Privacy Aware Learning (2012, background): “Q ( · Let ) denote the marginal probability deﬁned by · | The mutual information between X and Z is the expected Kullback-Leibler (KL) divergence between Q ( | X ) and Q ( ) : We view the problem of privacy as a game between the adversary controlling P and the data owners, who use Q to obscure the s”
Learning in a Large Function Space: Privacy-Preserving Mechanisms for SVM Learning (2009, positive): “The connection between algorithmic stability and di erential privacy is qualitatively similar to the recent work of Dwork and Lei (2009) who demonstrated that robust estimators can serve as the basis for private mechanisms, by exploiting the limited in uence of outliers on such estimators.”
Learning Differentially Private Language Models Without Losing Accuracy (2017, background): “The main goal of this work is to provide a strong guarantee that the trained model protects the privacy of individuals’ data without undue sacriﬁce in model quality, building upon recent work in applying differential privacy to machine learning [11, 10, 16, 8, 6, 5, 18, 24, 22, 1, 20, 26].”
Distributed Private Data Analysis: On Simultaneously Solving How and What (2008, background): “It has been demonstrated that differentially private analyses exist for a variety of tasks including the approximation of numerical functions (by adding carefully chosen random noise that conceals any single individual’s contribution) [11, 2, 22, 17], non-numerical analyses [20], datamining [2, 22],”
Minimax Optimal Procedures for Locally Private Estimation (2016, background): “More recent work aims to link differential privacy to statistical concerns (Dwork and Lei 2009; Wasserman and Zhou 2010; Hall, Rinaldo, and Wasserman 2012; Smith 2011; Chaudhuri, Monteleoni, and Sarwate 2011; Rubinstein et al. 2012); in particular, researchers have developed algorithms for private…”