Linear Program Reconstruction in Practice

TL;DR

This paper demonstrates a practical linear program reconstruction attack on a real statistical queries system, empirically evaluating its effectiveness and comparing it with related algorithms under various conditions.

Contribution

It presents the first practical implementation of a linear program reconstruction attack on a real dataset, assessing its performance and robustness.

Findings

Successful attack on production system

Effectiveness varies with dataset size and noise levels

Comparison shows strengths and limitations of algorithms

Abstract

We briefly report on a successful linear program reconstruction attack performed on a production statistical queries system and using a real dataset. The attack was deployed in test environment in the course of the Aircloak Challenge bug bounty program and is based on the reconstruction algorithm of Dwork, McSherry, and Talwar. We empirically evaluate the effectiveness of the algorithm and a related algorithm by Dinur and Nissim with various dataset sizes, error rates, and numbers of queries in a Gaussian noise setting.