Learning Differentially Private Mechanisms

TL;DR

This paper introduces an automated method for transforming non-private programs into differentially private algorithms by combining example-based input selection, continuous optimization, and symbolic mapping, improving accuracy and reliability.

Contribution

It presents a novel technique that automates the synthesis of differentially private algorithms from non-private code, addressing a key challenge in privacy-preserving data analysis.

Findings

Successfully learns foundational differential privacy algorithms

Outperforms baseline program synthesis methods

Provides a scalable approach for privacy-preserving algorithm design

Abstract

Differential privacy is a formal, mathematical definition of data privacy that has gained traction in academia, industry, and government. The task of correctly constructing differentially private algorithms is non-trivial, and mistakes have been made in foundational algorithms. Currently, there is no automated support for converting an existing, non-private program into a differentially private version. In this paper, we propose a technique for automatically learning an accurate and differentially private version of a given non-private program. We show how to solve this difficult program synthesis problem via a combination of techniques: carefully picking representative example inputs, reducing the problem to continuous optimization, and mapping the results back to symbolic expressions. We demonstrate that our approach is able to learn foundational algorithms from the differential privacy literature and significantly outperforms natural program synthesis baselines.