Code and data for Linear Program Reconstruction in Practice
Description
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.
The files in ./data were generated using the banking.psql database available at https://download.aircloak.com/ (last retrieved 12 Dec 2018). The best places to get started ./code/attack_script.m and ./code/simulated_experiment_script.m.
Notes
Files
journalprivacyconfidentiality/Linear-Program-Reconstruction-v20200114.zip
Files
(1.3 MB)
Name | Size | Download all |
---|---|---|
md5:c17622e96539900677d291b38ab4c29d
|
1.3 MB | Preview Download |
Additional details
Related works
- Is cited by
- Journal article: 10.29012/jpc.711 (DOI)
- Is supplement to
- https://github.com/journalprivacyconfidentiality/Linear-Program-Reconstruction/tree/v20200114 (URL)
- https://github.com/a785236/Linear-Program-Reconstruction (URL)
- Journal article: 10.29012/jpc.711 (DOI)
Funding
- TWC: TTP Option: Frontier: Collaborative: MACS: A Modular Approach to Cloud Security 1413920
- National Science Foundation