Differentially Private Synthetic Data via APIs 3: Using Simulators Instead of Foundation Model

TL;DR

This paper introduces Sim-PE, a method that uses simulators instead of foundation models within the Private Evolution framework to generate differentially private synthetic data, broadening applicability and improving performance.

Contribution

It demonstrates that simulators, including non-neural network data synthesizers, can be integrated into PE, expanding its use beyond foundation models for DP data synthesis.

Findings

Sim-PE improves classification accuracy by up to 3x.

Reduces FID by up to 80%.

Offers greater efficiency in DP synthetic data generation.

Abstract

Differentially private (DP) synthetic data, which closely resembles the original private data while maintaining strong privacy guarantees, has become a key tool for unlocking the value of private data without compromising privacy. Recently, Private Evolution (PE) has emerged as a promising method for generating DP synthetic data. Unlike other training-based approaches, PE only requires access to inference APIs from foundation models, enabling it to harness the power of state-of-the-art (SoTA) models. However, a suitable foundation model for a specific private data domain is not always available. In this paper, we discover that the PE framework is sufficiently general to allow APIs beyond foundation models. In particular, we demonstrate that many SoTA data synthesizers that do not rely on neural networks--such as computer graphics-based image generators, which we refer to as simulators--can be effectively integrated into PE. This insight significantly broadens PE's applicability and unlocks the potential of powerful simulators for DP data synthesis. We explore this approach, named Sim-PE, in the context of image synthesis. Across four diverse simulators, Sim-PE performs well, improving the downstream classification accuracy of PE by up to 3x, reducing FID by up to 80%, and offering much greater efficiency. We also show that simulators and foundation models can be easily leveraged together within PE to achieve further improvements. The code is open-sourced in the Private Evolution Python library: https://github.com/microsoft/DPSDA.