Class Attribute Inference Attacks: Inferring Sensitive Class Information by Diffusion-Based Attribute Manipulations

Lukas Struppek, Dominik Hintersdorf, Felix Friedrich, Manuel Brack, Patrick Schramowski +1 more
3/16/2023
cs.LGcs.CRcs.CV

No official code found. Best implementation: LukasStruppek/Class_Attribute_Inference_Attacks (5 stars, MIT, PyTorch, last commit 2024-10-25).

Abstract

Neural network-based image classifiers are powerful tools for computer vision tasks, but they inadvertently reveal sensitive attribute information about their classes, raising concerns about their privacy. To investigate this privacy leakage, we introduce the first Class Attribute Inference Attack (CAIA), which leverages recent advances in text-to-image synthesis to infer sensitive attributes of individual classes in a black-box setting, while remaining competitive with related white-box attacks. Our extensive experiments in the face recognition domain show that CAIA can accurately infer undisclosed sensitive attributes, such as an individual's hair color, gender, and racial appearance, which are not part of the training labels. Interestingly, we demonstrate that adversarial robust models are even more vulnerable to such privacy leakage than standard models, indicating that a trade-off between robustness and privacy exists.

View on arXivView PDF

Code Implementations(1)

LukasStruppek/Class_Attribute_Inference_AttacksCommunity100%
PyTorchMIT

Source code for our paper "Image Classifiers Leak Sensitive Attributes About Their Classes".

51Python, DockerfileMar 16, 20231 years agoMIT
computer-visiondeep-learningprivacypytorch

Similar Papers

Privacy Preserving Properties of Vision Classifiers

85% match2025

Inference Attacks Against Face Recognition Model without Classification Layers

85% match2024

Private Attribute Inference from Images with Vision-Language Models

84% match1 repo212024

Undermining Image and Text Classification Algorithms Using Adversarial Attacks

84% match2024

Only My Model On My Data: A Privacy Preserving Approach Protecting one Model and Deceiving Unauthorized Black-Box Models

84% match2024

Evading Forensic Classifiers with Attribute-Conditioned Adversarial Faces

84% match1 repo462023

Cite this paper

@article{struppek2023class,
  title  = {Class Attribute Inference Attacks: Inferring Sensitive Class Information by Diffusion-Based Attribute Manipulations},
  author = {Lukas Struppek and Dominik Hintersdorf and Felix Friedrich and Manuel Brack and Patrick Schramowski and Kristian Kersting},
  year   = {2023},
  eprint = {2303.09289},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2303.09289v2}
}

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/2303.09289.svg)](https://codeofpaper.com/paper/2303.09289)

Discussion