CS 369
A Study of Perturbation Techniques for Data Privacy

Instructors:   Cynthia Dwork and Nina Mishra and Kobbi Nissim
Class Schedule: Tuesdays, 1-3:00pm, Gates 259
Office Hours:   email to schedule
Number of Units: 2

Course Overview

The digital age has enabled widespread access to and collection of data. While there are several advantages to ubiquitous access to data, there is also the potential for breaching the privacy of individuals.

In a statistical database, personal information about n individuals is typically stored (n is usually very large). A statistical database system gives users the ability to obtain aggregate statistical information (like medians, averages, counts) and yet also preserve the privacy of individuals. Typical applications include medical, financial, and census data.

The course will study techniques for simultaneously enabling access to aggregate data and preserving privacy. Data perturbation is a classical technique for solving this problem. There are two flavors of data perturbation. In one version, the data are perturbed once, and the perturbed values are published. In the second version, the data are held secret; the database algorithm computes the true response to queries, and adds noise to the answer, reporting only the noisy answer. Both versions of the problem have a rich literaure.
The goal of the course is to

• understand the different definitions of data privacy
• understand the techniques for achieving privacy
• assess to what extent the suggested measures actually provide data privacy
• suggest new definitions of privacy
• suggest new algorithms for providing privacy
  

Background in probability, statistics, cryptography, and algorithms would be helpful.

Grading

The course can only be taken pass/fail and not for a letter grade. There will not be any exams.  The grade in the class will be based on a presentation done the last day of class (June 1st) on a paper selected from the list below. Please communicate which paper(s) you plan to present by May 27, 2004. If you'd like to discuss a paper that's not on the list, please email the instructors.

Topics

• Query Restriction/Auditing

• Input Perturbation

• Output Perturbation
  

Schedule

Reading List

Books/Surveys

• Security-Control Methods for Statistical Databases: A Comparative Study. N.R. Adam and J.C. Wortmann. ACM Computing Surveys 21(1989).
• Foundations of Cryptography: Volume 1, Basic Tools.  O. Goldreich.  2001.
  
• Foundations of Cryptography: Volume 2, Basic Applications.  O. Goldreich. 2004.
• Denning, Cryptography and Data Security, Addison-Wesley, 1982.

Applications

• Census Confidentiality and Privacy: 1790 - 2002. U.S. Bureau of the Census, 2002.
• The U.S. Census Bureau Respects Your Privacy and Keeps Your Personal Information Confidential. U.S. Bureau of the Census, 2001.

Cryptography
   Secure Function Evaluation

• Secure Computation: A tutorial.  Joe Killian.  Aladdin Workshop on Privacy in Data.  2003.
• 
  
• Cryptographic Techniques for Privacy-Preserving Data Mining. B. Pinkas. ACM SIGKDD Explorations 4 (2003).
• Privacy Preserving Data Mining, Y. Lindell and B. Pinkas, Journal of Cryptology 13 (2002).
• Secure Computation of the kth Ranked Element.  G. Aggarwal, N. Mishra, and B. Pinkas, Proc. of IACR Eurocrypt, 2004.
• Efficient Private Matching and Set Intersection.  M. Freedman, K. Nissim and B. Pinkas.  Proc. of IACR Eurocrypt, 2004.

   Semantic Security

• Lecture Notes in Cryptography.  M. Bellare and S. Goldwasser.  August 2001.
  

Query Restriction/Auditing

• Security in statistical databases for queries with small counts.  Chin.  ACM Transactions on Database Systems (TODS).  Volume 3 ,  Issue 1.  pages 92-104. 1978
• Secure databases: protection against user influence, Dobkin, Jones, Lipton, ACM Transactions on Database Systems (TODS), Volume 4 ,  Issue 1  (March 1979), 97 - 106.
• Security in Databases: A Combinatorial Study.  S. Reiss.  Journal of the ACM (JACM),  Volume 26 ,  Issue 1.
  
  Pages: 45 - 57.  1979.
  
• Auditing for secure statistical databases. F. Chin and G. Ozsoyoglu. ACM 81 Conference (1981).
• Security problems on inference control for SUM, MAX, and MIN queries.  F. Chin.  Journal of the ACM (JACM).  Volume 33 ,  Issue 3.   Pages: 451-464.  1986.
• Auditing boolean attributes, Kleinberg, Papadimitriou, Raghavan, Journal of Computer and System Sciences, Volume 66 ,  Issue 1  (February 2003), Special issue on PODS 2000, Pages: 244 - 253, 2003.

Sampling Contingency Tables

• Algebraic Algorithms for Sampling from Conditional Distributions.  P. Diaconis and B. Sturmfels.  Ann. Statist.  26, 363-97 (1998)
• Groebner Bases and Convex Polytopes, B. Sturmfels.  University Lecture Notes Series Vol 8, AMS 1995
  

Cell Suppression

• Suppression methodology and statistical disclosure control. L.H. Cox. JASA 75 (1980).
• A graph theoretic approach to satistical data security. D. Gusfield. SIAM J. Comput., 75:552--571, 1989.
• Data security equals graph connectivity, Ming-Yang Kao. SIAM Journal on Discrete Mathematics, Volume 9, Number 1,pp. 87-100
• Comparison of Systems Implementing Automated Cell Suppression for Economic Statistics, Nancy Kirkendall and Gordon Sande, Journal of Official Statistics, Vol.14, No.4, 1998.
• Network models for complementary cell suppression.  L. Cox.  Journal of the American Statistical Association, 90 1453-1462.
• Models and Methods for the Microdata Protection Problem, C.A.J. Hurkens and S.R. Tiourine.  Journal of Official Statistics, Vol.14, No.4, 1998.

Data Swapping

• Practical data-swapping: the first steps.   S. Reiss.  ACM Transactions on Database Systems (TODS),
  Volume 9 ,  Issue 1.  Pages 20-37.  1984
• Controlled data-swapping techniques for masking public use microdata sets. R.A. Moore, Jr.  SRD Report RR 96-04, U.S. Bureau of the Census, 1996.

Input Perturbation

• Masking Microdata Files.  J. Kim and W. Winkler.  Bureau of the Census.  1997.
• Multiplicative Noise for Masking Continuous Data.  J. Kim and W. Winkler.  2003.
  
• Disclosure Limitation Using Perturbation and Related Methods for Categorical Data
• Stephen E. Fienberg, Udi E. Makov, and Russell J. Steele. Journal of Official Statistics, Vol.14, No.4, 1998.
• Preserving Confidentiality AND Providing Adequate Data for Statistical Modeling: The Role of Partial and Perturbed Data (abstract) (ppt) (pdf) S. Fienberg. ALADDIN Workshop on Privacy in Data, 2003.
• Protecting data through 'Perturbation' Techniques: Impact on the knowledge discovery process.  R.L. Wilson and P. A. Rosen. Journal of Database Management, 14(2), 14-26, April-June 2003.
• Security of Random Data Perturbation Methods. K. Muralidhar and R. Sarathy. ACM TODS 24 (1999).
• Privacy-Preserving Data Mining, R. Agrawal and R. Srikant. SIGMOD 2000.
• Application and Analysis of the Mixture-of-Normals Approach to Masking Census Public-use Microdata.  G. Roque.  Manuscript.
  
• On the Design and Quantification of Privacy Preserving Data Mining Algorithms. D. Agrawal and C.C. Aggarwal. PODS 2001.
• Limiting Privacy Breaches in Privacy Preserving Data Mining. A. Evfimievski, J. Gehrke, and R. Srikant. PODS 2003.

Output Perturbation

• Revealing Information while Preserving Privacy. I. Dinur and K. Nissim, PODS 2003.  (abstract) (ppt) (html) (pdf)

• Privacy Preserving Data Mining on Vertically Partitioned Databases.  C. Dwork and K. Nissim.  Manuscript.  2004.

• Towards Privacy in Public Databases. S. Chawla, C. Dwork, F. McSherry, A. Smith, H. Wee. Manuscript.

Synthetic Datasets

• Discussion: Statistical Disclosure Limitation, Donald B. Rubin,Journal of Official Statistics Vol 9 No 2, 1993, pp. 461-468
• Multiple Imputation for Statistical Disclosure Limitation, T. E. Raghunathank, J.P. Reiter, and D. B. Rubin, JoS 19 (1), 2003, pp 1- 16
• Satisfying Disclosure Restrictions with Synthetic Data Sets, Jerome P. Reiter, JoS 18 (4), 2002, pp. 531-543
• Inference for Partially Synthetic, Public Use Microdata, J.P. Reiter, Survey Methodology 29 (2), 2003, pp. 181-188.