Subsampled Rényi Differential Privacy and Analytical Moments Accountant

We study the problem of subsampling in differential privacy (DP), a question that is the centerpiece behind many successful differentially private machine learning algorithms. Specifically, we provide a tight upper bound on the Renyi Differential Privacy (RDP) (Mironov, 2017) parameters for algorithms that: (1) subsample the dataset, and then (2) applies a randomized mechanism M to the subsample, in terms of the RDP parameters of M and the subsampling probability parameter. Our results generalize the moments accounting technique, developed by Abadi et al. (2016) for the Gaussian mechanism, to any subsampled RDP mechanism.

[1]  Yu-Xiang Wang,et al.  Per-instance Differential Privacy , 2017, J. Priv. Confidentiality.

[2]  Ilya Mironov,et al.  Rényi Differential Privacy , 2017, 2017 IEEE 30th Computer Security Foundations Symposium (CSF).

[3]  Thomas Steinke,et al.  Concentrated Differential Privacy: Simplifications, Extensions, and Lower Bounds , 2016, TCC.

[4]  Yu-Xiang Wang,et al.  Improving the Gaussian Mechanism for Differential Privacy: Analytical Calibration and Optimal Denoising , 2018, ICML.

[5]  Yu-Xiang Wang.  Per-instance Differential Privacy and the Adaptivity of Posterior Sampling in Linear and Ridge regression , 2017, ArXiv.

[6]  Amos Beimel,et al.  Characterizing the sample complexity of private learners , 2013, ITCS '13.

[7]  James R. Foulds,et al.  On the Theory and Practice of Privacy-Preserving Bayesian Data Analysis , 2016, UAI.

[8]  Ian Goodfellow,et al.  Deep Learning with Differential Privacy , 2016, CCS.

[9]  Peter Harremoës,et al.  Rényi Divergence and Kullback-Leibler Divergence , 2012, IEEE Transactions on Information Theory.

[10]  Moni Naor,et al.  Our Data, Ourselves: Privacy Via Distributed Noise Generation , 2006, EUROCRYPT.

[11]  Stephen E. Fienberg,et al.  Learning with Differential Privacy: Stability, Learnability and the Sufficiency and Necessity of ERM Principle , 2015, J. Mach. Learn. Res..

[12]  Aaron Roth,et al.  The Algorithmic Foundations of Differential Privacy , 2014, Found. Trends Theor. Comput. Sci..

[13]  Guy N. Rothblum,et al.  Boosting and Differential Privacy , 2010, 2010 IEEE 51st Annual Symposium on Foundations of Computer Science.

[14]  Fady Alajaji,et al.  Rényi divergence measures for commonly used univariate continuous distributions , 2013, Inf. Sci..

[15]  Kobbi Nissim,et al.  Differentially Private Release and Learning of Threshold Functions , 2015, 2015 IEEE 56th Annual Symposium on Foundations of Computer Science.

[16]  Raef Bassily,et al.  Differentially Private Empirical Risk Minimization: Efficient Algorithms and Tight Error Bounds , 2014, 1405.7085.

[17]  Miljenko Huzak,et al.  Characteristic Functions , 2011, International Encyclopedia of Statistical Science.

[18]  Guy N. Rothblum,et al.  Concentrated Differential Privacy , 2016, ArXiv.

[19]  Frank Nielsen,et al.  On the chi square and higher-order chi distances for approximating f-divergences , 2013, IEEE Signal Processing Letters.

[20]  A. Ihler,et al.  On the Theory and Practice of Privacy-Preserving Bayesian Data Analysis , 2016 .

[21]  L. Sweeney.  Only You, Your Doctor, and Many Others May Know , 2015 .

[22]  Alexander J. Smola,et al.  Privacy for Free: Posterior Sampling and Stochastic Gradient Monte Carlo , 2015, ICML.

[23]  Salil P. Vadhan,et al.  The Complexity of Computing the Optimal Composition of Differential Privacy , 2015, IACR Cryptol. ePrint Arch..

[24]  Cynthia Dwork,et al.  Calibrating Noise to Sensitivity in Private Data Analysis , 2006, TCC.

[25]  Anand D. Sarwate,et al.  Stochastic gradient descent with differentially private updates , 2013, 2013 IEEE Global Conference on Signal and Information Processing.

[26]  Pramod Viswanath,et al.  The Composition Theorem for Differential Privacy , 2013, IEEE Transactions on Information Theory.

[27]  S. G. Bobkov,et al.  Rényi divergence and the central limit theorem , 2016, The Annals of Probability.

[28]  Kamalika Chaudhuri,et al.  Renyi Differential Privacy Mechanisms for Posterior Sampling , 2017, NIPS.

[29]  Úlfar Erlingsson,et al.  RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response , 2014, CCS.

[30]  Thomas Steinke,et al.  Composable and versatile privacy via truncated CDP , 2018, STOC.

[31]  William G. Madow,et al.  On the Limiting Distributions of Estimates Based on Samples from Finite Universes , 1948 .

[32]  Sofya Raskhodnikova,et al.  What Can We Learn Privately? , 2008, 2008 49th Annual IEEE Symposium on Foundations of Computer Science.

[33]  Gilles Barthe,et al.  Privacy Amplification by Subsampling: Tight Analyses via Couplings and Divergences , 2018, NeurIPS.