The Geometry of Differential Privacy: the Sparse and Approximate Cases

In this work, we study trade-offs between accuracy and privacy in the context of linear queries over histograms. This is a rich class of queries that includes contingency tables and range queries, and has been a focus of a long line of work. For a set of d linear queries over a database x N, we seek to find the differentially private mechanism that has the minimum mean squared error. For pure differential privacy, an O (² d) approximation to the optimal mechanism is known. Our first contribution is to give an O (² d) approximation guarantee for the case of (, δ) -differential privacy. Our mechanism is simple, efficient and adds correlated Gaussian noise to the answers. We prove its approximation guarantee relative to the hereditary discrepancy lower bound of Muthukrishnan and Nikolov, using tools from convex geometry. We next consider this question in the case when the number of queries exceeds the number of individuals in the database, i. e. when d > n \|x\|₁. It is known that better mechanisms exist in this setting. Our second main contribution is to give an (, δ) -differentially private mechanism which is optimal up to a (d, N) factor for any given query set A and any given upper bound n on \|x\|₁. This approximation is achieved by coupling the Gaussian noise addition approach with a linear regression step. We give an analogous result for the -differential privacy setting. We also improve on the mean squared error upper bound for answering counting queries on a database of size n by Blum, Ligett, and Roth, and match the lower bound implied by the work of Dinur and Nissim up to logarithmic factors. The connection between hereditary discrepancy and the privacy mechanism enables us to derive the first polylogarithmic approximation to the hereditary discrepancy of a matrix A.