Abstract
Randomized response is attractive for privacy preserving data collection because the provided privacy can be quantified by means such as differential privacy. However, recovering and analyzing statistics involving multiple dependent randomized binary attributes can be difficult, posing a significant barrier to use. In this work, we address this problem by identifying and analyzing a family of response randomizers that change each binary attribute independently with the same probability. Modes of Google’s Rappor randomizer as well as applications of two well-known classical randomized response methods, Warner’s original method and Simmons’ unrelated question method, belong to this family. We show that randomizers in this family transform multinomial distribution parameters by an iterated Kronecker product of an invertible and bisymmetric \(2\times 2\) matrix. This allows us to present a simple and efficient algorithm for obtaining unbiased maximum likelihood parameter estimates for \(k\)-way marginals from randomized responses and provide theoretical bounds on the statistical efficiency achieved. We also describe the efficiency – differential privacy tradeoff. Importantly, both randomization of responses and the estimation algorithm are simple to implement, an aspect critical to technologies for privacy protection and security.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Distributed as the flat Dirichlet distribution of order \(2^n\).
References
Abul-Ela, A.L.A., Greenberg, B.G., Horvitz, D.G.: A multi-proportions randomized response model. J. Am. Stat. Assoc. 62(319), 990–1008 (1967). https://doi.org/10.2307/2283687
Apple: Learning with Privacy at Scale - Apple. December 2017. https://machinelearning.apple.com/2017/12/06/learning-with-privacy-at-scale.html
Barabesi, L., Franceschi, S., Marcheselli, M.: A randomized response procedure for multiple-sensitive questions. Stat. Papers 53(3), 703–718 (2012). https://doi.org/10.1007/s00362-011-0374-5
Blair, G., Imai, K., Zhou, Y.Y.: Design and analysis of the randomized response technique. J. Am. Stat. Assoc. 110(511), 1304–1319 (2015). https://doi.org/10.1080/01621459.2015.1050028
Bourke, P.D.: Randomized response multivariate designs for categorical data. Commun. Stat. Theory Methods 11(25), 2889–2901 (1982). https://doi.org/10.1080/03610928208828430
Casella, G., Berger, R.L.: Statistical Inference. Duxbury/Thomson Learning, Australia; Pacific Grove, CA (2002)
Dalenius, T.: Towards a methodology for statistical disclosure control. Statistisk Tidskrift 15(429–444), 2–1 (1977)
Duchi, J.C., Jordan, M.I., Wainwright, M.J.: Local privacy and statistical minimax rates. In: 2013 51st Annual Allerton Conference on Communication, Control, and Computing (Allerton), pp. 1592–1592, October 2013. https://doi.org/10.1109/Allerton.2013.6736718
Dwork, C., McSherry, F., Nissim, K., Smith, A.: Calibrating noise to sensitivity in private data analysis. In: Proceedings of the Conference on Theory of Cryptography (2006). https://doi.org/10.1007/11681878%5F14
Erlingsson, Ú., Pihur, V., Korolova, A.: RAPPOR: Randomized aggregatable privacy-preserving ordinal response. In: Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security, pp. 1054–1067. CCS 2014, ACM, New York (2014). https://doi.org/10.1145/2660267.2660348
Fanti, G., Pihur, V., Erlingsson, Ú.: Building a RAPPOR with the unknown: privacy-preserving learning of associations and data dictionaries. arXiv:1503.01214 [cs], March 2015. http://arxiv.org/abs/1503.01214
Folsom, R.E., Greenberg, B.G., Horvitz, D.G., Abernathy, J.R.: The two alternate questions randomized response model for human surveys. J. Am. Stat. Assoc. 68(343), 525–530 (1973). https://doi.org/10.2307/2284771
Greenberg, B.G., Abul-Ela, A.L.A., Simmons, W.R., Horvitz, D.G.: The unrelated question randomized response model: theoretical framework. J. Am. Stat. Assoc. 64(326), 520–539 (1969). https://doi.org/10.2307/2283636
Kairouz, P., Bonawitz, K., Ramage, D.: Discrete distribution estimation under local privacy. In: Proceedings of the 33rd International Conference on International Conference on Machine Learning, vol. 48, pp. 2436–2444. ICML 2016, JMLR.org, New York (2016). http://dl.acm.org/citation.cfm?id=3045390.3045647
Kairouz, P., Oh, S., Viswanath, P.: Extremal mechanisms for local differential privacy. In: Ghahramani, Z., Welling, M., Cortes, C., Lawrence, N.D., Weinberger, K.Q. (eds.) Advances in Neural Information Processing Systems, vol. 27, pp. 2879–2887. Curran Associates, Inc. (2014). http://papers.nips.cc/paper/5392-extremal-mechanisms-for-local-differential-privacy.pdf
Lehmann, E.L., Casella, G.: Theory of Point Estimation. STS. Springer, New York (1998). https://doi.org/10.1007/b98854
Lensvelt-Mulders, G.J.L.M., Hox, J.J., van der Heijden, P.G.M., Maas, C.J.M.: Meta-analysis of randomized response research: thirty-five years of validation. Sociol. Methods Res. 33(3), 319–348 (2005). https://doi.org/10.1177/0049124104268664
Moran, P.A.P.: The random division of an interval. Suppl. J. Roy. Stat. Soc. 9(1), 92–98 (1947). https://doi.org/10.2307/2983572
Müller-Hannemann, M., Schirra, S. (eds.): Algorithm Engineering. LNCS, vol. 5971. Springer, Heidelberg (2010). https://doi.org/10.1007/978-3-642-14866-8
SOSA: Symposium on Simplicity in Algorithms, January 2018. https://simplicityalgorithms.wixsite.com/sosa/cfp
Tang, J., Korolova, A., Bai, X., Wang, X., Wang, X.: Privacy loss in apple’s implementation of differential privacy on MacOS 10.12. arXiv:1709.02753 [cs], September 2017. http://arxiv.org/abs/1709.02753
Umesh, U.N., Peterson, R.A.: A critical evaluation of the randomized response method: applications, validation, and research agenda. Sociol. Methods Res. 20(1), 104–138 (1991). https://doi.org/10.1177/0049124191020001004
Warner, S.L.: Randomized Response: a survey technique for eliminating evasive answer bias. J. Am. Stat. Assoc. 60(309), 63–69 (1965). https://doi.org/10.1080/01621459.1965.10480775
Acknowledgments
Thanks go to the anonymous reviewers for their comments. This work was in part funded by Oppland fylkeskommune.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
A Proofs
A Proofs
We start by making a key observation.
Observation 1: Consider the \(2^n \times 2^n\) matrix \(C\). If we let entry \(C_{ix', jy'} = \eta ((ix') \oplus (jy')) = 2^{\eta (i \oplus j)} \;\eta (x' \oplus y'),\) we get that \(C_{x,y} = 1^{n - |x \oplus y|} \; 2^{|x \oplus y|}\). Since we can write \(C = J_1 \otimes J_2 \otimes \cdots \otimes J_n\) where \(J_k\) is the \(2 \times 2\) matrix \(J\) such that \(J_{i,j} = 2^{i \oplus j}\), i.e., \(J = \begin{pmatrix} 1 &{} 2 \\ 2 &{} 1 \end{pmatrix},\) we get that \(D_{x,y} = a^{n - |x \oplus y|} \; b^{|x \oplus y|}\) for \(D = C_{a,b}(n) = B_1 \otimes B_2 \otimes \cdots \otimes B_n\) where \(B_k = \begin{pmatrix} a &{} b \\ b &{} a \end{pmatrix}.\)
Proof of Proposition 1: Note that we can write
where \(F\) and \(G\) are functions independent of \(m\). Then for any positive integer \(m\),
and
Furthermore,
and similarly \({\text {E}}\left( \Vert \hat{\pi }^*(m) - \pi \Vert _2^2\right) = \sum _x{\text {Var}}(\hat{\pi }^*_x(m))\). \(\square \)
Proof of Proposition 2: The proposition follows directly from Theorem 1. \(\square \)
Proof of Proposition 3: We first note that for \(i \in \{0, 1, \ldots , 2^n - 1\}\) we have that \(\varsigma ((2^n - 1) - i) = \mathbf {1} \oplus \varsigma (i)\). From this and that \(\oplus \) commutes, we get
-
1.
\(|\varsigma (i) \oplus \varsigma (n - i)| = n\), and
-
2.
\(|\varsigma (i) \oplus \varsigma (j)| = |\varsigma (2^n - 1 - j) \oplus \varsigma (2^n - 1 - i)|\).
The above and that the entry \(C_{a,b}(n)_{i,j} = g(|\varsigma (i) \oplus \varsigma (j)|, a, b)\) for some \(g\), the proposition follows. \(\square \)
Proof of Theorem 2: The first equation follows directly from Observation 1. We have that \(C_{a,b}(1)\) is invertible if \(a^2 \ne b^2\). From this and that \((A \otimes B) = (A^{-1} \otimes B^{-1})\) we complete the proof. \(\square \)
Proof of Lemma 1: From Sect. 3.2 we have that
By properties of the trace of matrix products and symmetry of \(C^{-1}\),
From \((A \otimes B)(C \otimes D) = (AC) \otimes (BD)\) it follows that \(C_{a}(n) C_{a}(n) = C_{a^2 + (1-a)^2}(n)\). From this and Theorem 2 and Corollary 1 we get that the entry \((C^{-1}C^{-1})_{0,0} = f(n, a)\) where
Furthermore, from Proposition 3 the diagonal entries of \(C^{-1}C^{-1}\) are all \(f(n,a)\). Combining this, that \({\text {Tr}}(AB) = \sum _{i,j}(A \odot B^T)_{i,j}\), and \(\sum _x C_x \pi = 1\),
and consequently, \({\text {Tr}}({\text {cov}}(\hat{\pi }_x(m))) = m^{-1} \left( f(n,a) - s\right) \). \(\square \)
Proof of Theorem 3: We have that
From Lemma 1 and Proposition 1 we get that \(L= f_L(s) = \frac{c - s}{(1 - s)}\) for \(c = \left( \frac{a^2+(1-a)^2}{(2a-1)^2}\right) ^n.\) From \(0 \le p_x \le 1\) and \(\sum _x p_x = 1\), \(s\) has a minimum when \(p_x = 1/2^n\) for all \(x\), and maximum when \(p_x = 1\) for a fixed \(x\), and \(p_y = 0\) for \(y \ne x\). These values are then \(\frac{2^n}{(2^n)^2} = 2^{-n}\) and \(1\), respectively. The \(m\)’th derivative of \(f_L(s) = \frac{c-s}{1-s}\) wrt. \(0 \le s < 1\) is \(f_L^{(m)}(s) = \frac{m!}{(1-s)^m} \left( f_L(s) - 1 \right) \). The loss \(f_L\) therefore achieves its minimum at \(f_L(2^{-n})\). \(\square \)
Proof of Proposition 4 (sketch): The \(m\)’th derivative of \(f_L(s) = \frac{c-s}{1-s}\) wrt. \(0 \le s < 1\) is \(f_L^{(m)}(s) = \frac{m!}{(1-s)^m} \left( f_L(s) - 1 \right) \). Since \(c \ge 1\), \(f^{(m)}_L \ge 0\) for all \(m > 0\). In particular, we have that \(f_L\) is convex, as is \(f_L^{(m)}\) for all \(m\). Using the expectation for a first order Taylor approximation for convex \(f_L\) we have that for random variable \(S\)
where \(\lambda = \max _{x \in \mathcal {I}} f^{(2)}_L(x) \ge 0\) for suitable interval \(\mathcal {I}\). Dividing \((*)\) by \(f_L({\text {E}}(S)) = L(n)\), we get
where
Let \(S = \pi ^T\pi \). Recalling that \(c = c(a)^n\) and expanding both numerator and denominator at \(n=3\) (where the minimum occurs since \(f_L\) is increasing and \({\text {Var}}(S)\) and \({\text {E}}(S)\) are both decreasing in \(n\)), we see that \(\delta (n) \in O(2^{-3n})\). Applying Chebyshev’s inequality, we have that \(P(S \ge {\text {E}}(S) + 10 {\text {Var}}(S)^{\frac{1}{2}}) \le 0.01\). Evaluating \(\delta \) at \({\text {E}}(S) + 10 {\text {Var}}(S)^{\frac{1}{2}}\) and \(n=3\), we arrive at the numerical bound. \(\square \)
Proof of Proposition 5: Let the computation of \(Z \otimes R\) require \(t_f(n^2)\) time for \(2\times 2\) matrix \(Z\) and \(R\) of size \(n \times n\). Then we can compute \(R_{a,b}(n)\) at a time cost of \(t(n) = t_f(2^{2(n-1)}) + t(n-1) = \sum _{i = 0}^n t_f(2^{2i}) = \sum _{i = 0}^n t_f(4^i).\) Letting \(t_f(n) = k 4 n\) for some \(k\), then \(t(n) = 4k\sum _{i=0}^n 2^{2i} = 4k \sum _{i=0}^n 4^{i} = 4k (1 + \frac{1-4^n}{1-4}) = 4k (1 + \frac{4^n-1}{3}).\) Now we have that \(t(n) = O(4^n) = O({2^n}^2) = O(|R_{a,b}(n)|)\). In other words, the singly recursive algorithm \(R_{a,b}(n)\) is linear in the time in the number of elements of the output matrix as we can perform \(t_f\) in linear time in the size of input \(R\), in fact we can expect that the Kronecker product can be implemented with \(k \le 3\), due to reading, multiplication, and writing. \(\square \)
Rights and permissions
Copyright information
© 2018 Springer Nature Switzerland AG
About this paper
Cite this paper
Vinterbo, S.A. (2018). A Simple Algorithm for Estimating Distribution Parameters from \(n\)-Dimensional Randomized Binary Responses. In: Chen, L., Manulis, M., Schneider, S. (eds) Information Security. ISC 2018. Lecture Notes in Computer Science(), vol 11060. Springer, Cham. https://doi.org/10.1007/978-3-319-99136-8_11
Download citation
DOI: https://doi.org/10.1007/978-3-319-99136-8_11
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-99135-1
Online ISBN: 978-3-319-99136-8
eBook Packages: Computer ScienceComputer Science (R0)