Statistical Data Privacy: A Song of Privacy and Utility

To quantify trade-offs between increasing demand for open data sharing and concerns about sensitive information disclosure, statistical privacy (SDP) methodology analyzes release mechanisms that sanitize outputs based on confidential data. Two dominant frameworks exist: disclosure control (SDC) the more recent differential (DP). Despite framing differences, both SDC DP share same problems at their core. For inference problems, either we may design optimal associated estimators satisfy bounds risk measures, or adjust existing sanitized output to create new statistically valid estimators. Regardless of adjustment, in evaluating utility, inferences from mechanism require uncertainty quantification accounts effect sanitization introduces bias and/or variance. In this review, discuss foundations common DP, highlight major developments SDP, present exciting research private inference.