Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2402.09478 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909547609718784 |
|---|---|
| author | Liu, Sheng Wang, Zihan Chen, Yuxiao Lei, Qi |
| author_facet | Liu, Sheng Wang, Zihan Chen, Yuxiao Lei, Qi |
| contents | Reconstruction attacks and defenses are essential in understanding the data leakage problem in machine learning. However, prior work has centered around empirical observations of gradient inversion attacks, lacks theoretical grounding, and cannot disentangle the usefulness of defending methods from the computational limitation of attacking methods. In this work, we propose to view the problem as an inverse problem, enabling us to theoretically and systematically evaluate the data reconstruction attack. On various defense methods, we derived the algorithmic upper bound and the matching (in feature dimension and architecture dimension) information-theoretical lower bound on the reconstruction error for two-layer neural networks. To complement the theoretical results and investigate the utility-privacy trade-off, we defined a natural evaluation metric of the defense methods with similar utility loss among the strongest attacks. We further propose a strong reconstruction attack that helps update some previous understanding of the strength of defense methods under our proposed evaluation metric. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_09478 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Data Reconstruction Attacks and Defenses: A Systematic Evaluation Liu, Sheng Wang, Zihan Chen, Yuxiao Lei, Qi Cryptography and Security Machine Learning Reconstruction attacks and defenses are essential in understanding the data leakage problem in machine learning. However, prior work has centered around empirical observations of gradient inversion attacks, lacks theoretical grounding, and cannot disentangle the usefulness of defending methods from the computational limitation of attacking methods. In this work, we propose to view the problem as an inverse problem, enabling us to theoretically and systematically evaluate the data reconstruction attack. On various defense methods, we derived the algorithmic upper bound and the matching (in feature dimension and architecture dimension) information-theoretical lower bound on the reconstruction error for two-layer neural networks. To complement the theoretical results and investigate the utility-privacy trade-off, we defined a natural evaluation metric of the defense methods with similar utility loss among the strongest attacks. We further propose a strong reconstruction attack that helps update some previous understanding of the strength of defense methods under our proposed evaluation metric. |
| title | Data Reconstruction Attacks and Defenses: A Systematic Evaluation |
| topic | Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2402.09478 |