10 A Framework for Adaptive Di erential Privacy ( Extended Version )

Di erential privacy is a widely studied theory for analyzing sensitive data with a strong privacy guarantee— any change in an individual’s data can have only a small statistical e ect on the result—and a growing number of programming languages now support di erentially private data analysis. A common shortcoming of these languages is poor support for adaptivity. In practice, a data analyst rarely wants to run just one function over a sensitive database, nor even a predetermined sequence of functions with xed privacy parameters; rather, she wants to engage in an interaction where, at each step, both the choice of the next function and its privacy parameters are informed by the results of prior functions. Existing languages support this scenario using a simple composition theorem, which often gives rather loose bounds on the actual privacy cost of composite functions, substantially reducing how much computation can be performed within a given privacy budget. The theory of di erential privacy includes other theorems with much better bounds, but these have not yet been incorporated into programming languages. We propose a novel framework for adaptive composition that is elegant, practical, and implementable. It consists of a reformulation based on typed functional programming of the privacy lters of Rogers et al. (2016), together with a concrete realization of this framework in the design and implementation of a new language, called Adaptive Fuzz. Adaptive Fuzz transplants the core static type system of Fuzz (Haeberlen et al. 2011) to the adaptive setting by wrapping the Fuzz typechecker and runtime system in an outer adaptive layer, allowing Fuzz programs to be conveniently constructed and typechecked on the y. We describe an interpreter for Adaptive Fuzz and report results from two case studies demonstrating its e ectiveness for implementing common statistical algorithms over real data sets.

[1]  Philip Wadler,et al.  Linear Types can Change the World! , 1990, Programming Concepts and Methods.

[2]  Ramakrishnan Srikant,et al.  Privacy-preserving data mining , 2000, SIGMOD '00.

[3]  Latanya Sweeney,et al.  k-Anonymity: A Model for Protecting Privacy , 2002, Int. J. Uncertain. Fuzziness Knowl. Based Syst..

[4]  Alexandre V. Evfimievski,et al.  Privacy preserving mining of association rules , 2002, Inf. Syst..

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

[6]  Ashwin Machanavajjhala,et al.  l-Diversity: Privacy Beyond k-Anonymity , 2006, ICDE.

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

[8]  Ninghui Li,et al.  t-Closeness: Privacy Beyond k-Anonymity and l-Diversity , 2007, 2007 IEEE 23rd International Conference on Data Engineering.

[9]  Adam D. Smith,et al.  Composition attacks and auxiliary information in data privacy , 2008, KDD.

[10]  Vitaly Shmatikov,et al.  Robust De-anonymization of Large Sparse Datasets , 2008, 2008 IEEE Symposium on Security and Privacy (sp 2008).

[11]  Daniel Kifer,et al.  Attacks on privacy and deFinetti's theorem , 2009, SIGMOD Conference.

[12]  Frank McSherry,et al.  Privacy integrated queries: an extensible platform for privacy-preserving data analysis , 2009, SIGMOD Conference.

[13]  Cynthia Dwork,et al.  Differential privacy and robust statistics , 2009, STOC '09.

[14]  Ilya Mironov,et al.  Differentially private recommender systems , 2009 .

[15]  Vitaly Shmatikov,et al.  Airavat: Security and Privacy for MapReduce , 2010, NSDI.

[16]  Benjamin C. Pierce,et al.  Distance makes the types grow stronger: a calculus for differential privacy , 2010, ICFP '10.

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

[18]  Andreas Haeberlen,et al.  Differential Privacy Under Fire , 2011, USENIX Security Symposium.

[19]  Andreas Haeberlen,et al.  DJoin: differentially private join queries over distributed databases , 2012, OSDI 2012.

[20]  Gilles Barthe,et al.  Probabilistic relational reasoning for differential privacy , 2012, POPL '12.

[21]  Elaine Shi,et al.  GUPT: privacy preserving data analysis made easy , 2012, SIGMOD Conference.

[22]  Ilya Mironov,et al.  On significance of the least significant bits for differential privacy , 2012, CCS.

[23]  Vitaly Shmatikov,et al.  Privacy-preserving data exploration in genome-wide association studies , 2013, KDD.

[24]  Andreas Haeberlen,et al.  Linear dependent types for differential privacy , 2013, POPL.

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

[26]  Gilles Barthe,et al.  Higher-Order Approximate Relational Refinement Types for Mechanism Design and Differential Privacy , 2014, POPL.

[27]  Andreas Haeberlen,et al.  Differential Privacy: An Economic Method for Choosing Epsilon , 2014, 2014 IEEE 27th Computer Security Foundations Symposium.

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

[29]  Shiva Prasad Kasiviswanathan,et al.  On the 'Semantics' of Differential Privacy: A Bayesian Formulation , 2008, J. Priv. Confidentiality.

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

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

[32]  Sorin Lerner,et al.  On Subnormal Floating Point and Abnormal Timing , 2015, 2015 IEEE Symposium on Security and Privacy.

[33]  Ryan J. Tibshirani,et al.  A general framework for fast stagewise algorithms , 2014, J. Mach. Learn. Res..

[34]  Aaron Roth,et al.  Privacy Odometers and Filters: Pay-as-you-Go Composition , 2016, NIPS.

[35]  David Sands,et al.  Featherweight PINQ , 2015, J. Priv. Confidentiality.