Combining Public and Private Data

TL;DR

This paper introduces optimized mixed estimators for combining public and private data under differential privacy, demonstrating improved accuracy over existing methods through empirical evaluation.

Contribution

It presents novel mixed estimators for mean and median that effectively combine heterogeneous privacy data, outperforming previous approaches.

Findings

Mixed estimators reduce variance compared to baseline methods.

Experiments show improved accuracy in aggregate statistics estimation.

Proposed mechanisms outperform existing techniques in empirical tests.

Abstract

Differential privacy is widely adopted to provide provable privacy guarantees in data analysis. We consider the problem of combining public and private data (and, more generally, data with heterogeneous privacy needs) for estimating aggregate statistics. We introduce a mixed estimator of the mean optimized to minimize the variance. We argue that our mechanism is preferable to techniques that preserve the privacy of individuals by subsampling data proportionally to the privacy needs of users. Similarly, we present a mixed median estimator based on the exponential mechanism. We compare our mechanisms to the methods proposed in Jorgensen et al. [2015]. Our experiments provide empirical evidence that our mechanisms often outperform the baseline methods.