Learning Robust and Privacy-Preserving Representations via Information Theory

Binghui Zhang, Sayedeh Leila Noorbakhsh, Yun Dong, Yuan Hong, Binghui Wang
12/15/2024
cs.LGcs.CR

No official code found. Best implementation: ARPRL/ARPRL (3 stars, MIT, last commit 2024-12-15).

Abstract

Machine learning models are vulnerable to both security attacks (e.g., adversarial examples) and privacy attacks (e.g., private attribute inference). We take the first step to mitigate both the security and privacy attacks, and maintain task utility as well. Particularly, we propose an information-theoretic framework to achieve the goals through the lens of representation learning, i.e., learning representations that are robust to both adversarial examples and attribute inference adversaries. We also derive novel theoretical results under our framework, e.g., the inherent trade-off between adversarial robustness/utility and attribute privacy, and guaranteed attribute privacy leakage against attribute inference adversaries.

View on arXivView PDF

Code Implementations(1)

ARPRL/ARPRLCommunity100%
MIT
30PythonMay 17, 20231 years agoMIT

Similar Papers

Inf2Guard: An Information-Theoretic Framework for Learning Privacy-Preserving Representations against Inference Attacks

89% match1 repo82024

Learning Private Representations through Entropy-based Adversarial Training

87% match2025

Robust Privacy: Inference-Time Privacy through Certified Robustness

87% match2026

Differentially Private and Adversarially Robust Machine Learning: An Empirical Evaluation

86% match2024

Range Membership Inference Attacks

85% match1 repo7242024

Robust Representation Learning for Privacy-Preserving Machine Learning: A Multi-Objective Autoencoder Approach

85% match2023

Cite this paper

@article{zhang2024learning,
  title  = {Learning Robust and Privacy-Preserving Representations via Information Theory},
  author = {Binghui Zhang and Sayedeh Leila Noorbakhsh and Yun Dong and Yuan Hong and Binghui Wang},
  year   = {2024},
  eprint = {2412.11066},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2412.11066v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2412.11066.svg)](https://codeofpaper.com/paper/2412.11066)

Discussion