Soft voting reduces noise in policy-aware private data synthesis
Soft Voting for Policy-Aware Private Data Synthesis
Cryptography and SecurityDatabases
Summary
Differential privacy protects data by adding noise, but sometimes it adds more noise than needed. The authors study a method called Blowfish privacy that protects only certain sensitive changes, potentially allowing less noise. They find that existing methods using hard voting don't benefit from this because their sensitivity is dominated by edge cases. They propose a soft voting method that changes smoothly, allowing tighter sensitivity bounds and less noise under some privacy settings. This method works best when protecting numeric data and can be predicted using public data before spending privacy budget.
What this means in practice
- •For data privacy engineers: Improve synthetic data generation by reducing noise when applying customized privacy policies with numeric data attributes.
- •For healthcare data teams: Generate private synthetic patient datasets that respect policy constraints with less distortion under strong privacy requirements.
Authors
Yingge Hu, Gautham Ramesh Babu, Mostafa Milani
Abstract
Blowfish privacy relaxes differential privacy (DP) by protecting only the attribute-value substitutions a data owner specifies as edges of a policy graph. A sparser policy can reduce the noise required by a mechanism, but only when the released statistic changes less across protected substitutions than across arbitrary DP neighbors. We study this question for evolutionary, nearest-neighbor DP synthesizers such as Private Evolution (PE) and its tabular instantiation Tab-PE, which score private records against a candidate population and release a noisy vote histogram. Their hard vote is constant inside each candidate's decision region and jumps at its boundary. Its policy-specific sensitivity therefore equals the full worst-case value whenever at least one protected substitution crosses a boundary, regardless of how short that substitution is. Because every round we examined contained such a substitution, the policy graph gave no reduction in noise. We propose BF-Soft, a temperature-smoothed soft vote whose response changes gradually with distance. Its sensitivity has a tight closed-form bound in the policy graph's reach and the temperature, independent of the number of candidates, and the bound can be computed once before synthesis. It also predicts from the policy alone when policy-aware smoothing cannot substantially reduce noise: protecting a flat categorical or binary attribute drives the reach to its maximum. On real and synthetic datasets under narrow numeric policies, BF-Soft reduces error relative to hard voting at strong privacy budgets, while the advantage reverses at weaker budgets. A public-data pilot predicts when soft voting is beneficial without spending private budget.