Differential privacy via wavelet transforms

Abstract

Privacy preserving data publishing has attracted considerable research interest in recent years. Among the existing solutions, -differential privacy provides one of the strongest privacy guarantees. Existing data publishing methods that achievedifferential privacy, however, offer little data utility. In particular, if the output dataset is used to answer count queries, the noise in the query answers can be proportional to the number of tuples in the data, which renders the results useless.In this paper, we develop a data publishing technique that ensures -differential privacy while providing accurate answers for range-count queries, i.e., count queries where the predicate on each attribute is a range. The core of our solution is a framework that applies wavelet transforms on the data before adding noise to it. We present instantiations of the proposed framework for both ordinal and nominal data, and we provide a theoretical analysis on their privacy and utility guarantees. In an extensive experimental study on both real and synthetic data, we show the effectiveness and efficiency of our solution., 978-1-4244-5446-4/10/$26.00

Differentially Private Synthesization of Multi-Dimensional Data using Copula Functions

“…Table 2 shows the domain sizes of the datasets. For nominal attributes, we convert them to numeric attributes by imposing a total order on the domain of the attribute as in [39]. …”

“…We do not include this method in our experimental comparison due to its high computational complexity. Xiao et al [39] propose a Privelet method by applying a wavelet transform on the original histogram, then adding polylogarithmic noise to the transformed data. Cormode et al [10] developed a series of filtering and sampling techniques to obtain a compact summary of DP histograms.…”

“…Based on the discussion above, we will experimentally compare the proposed DPCopula method with the Privelet+ [39], Filter Priority (FP) method [10], PSD method [9], and P-HP method [1] as representatives of the general-purpose histogram methods.…”

“…There are several state-of-the-art techniques for obtaining one-dimensional DP histograms effectively and efficiently, such as PSD, Privelet [39], NoiseFirst and StructureFirst [41], EFPA [1]. Here we use EFPA to generate DP marginal histograms which is superior to other methods.…”

“…[5, 27, 39, 19, 12, 41, 9, 10]). Given an original dataset, the goal is to publish a DP statistical summary such as marginal or multi-dimensional histograms that can be used to answer predicate queries or to generate DP synthetic data that mimic the original data.…”