Unlearnable Examples: Making Personal Data Unexploitable
Hanxun Huang, Xingjun Ma, Sarah Monazam Erfani, James Bailey, Yisen, Wang

TL;DR
This paper introduces error-minimizing noise as a method to make personal data unlearnable by deep learning models, protecting privacy without compromising data utility.
Contribution
It proposes a novel error-minimizing noise technique that renders training data unexploitable while remaining imperceptible to humans.
Findings
Error-minimizing noise effectively prevents models from learning from protected data.
The method works for both sample-wise and class-wise data.
It is practical and effective in face recognition scenarios.
Abstract
The volume of "free" data on the internet has been key to the current success of deep learning. However, it also raises privacy concerns about the unauthorized exploitation of personal data for training commercial models. It is thus crucial to develop methods to prevent unauthorized data exploitation. This paper raises the question: \emph{can data be made unlearnable for deep learning models?} We present a type of \emph{error-minimizing} noise that can indeed make training examples unlearnable. Error-minimizing noise is intentionally generated to reduce the error of one or more of the training example(s) close to zero, which can trick the model into believing there is "nothing" to learn from these example(s). The noise is restricted to be imperceptible to human eyes, and thus does not affect normal data utility. We empirically verify the effectiveness of error-minimizing noise in both…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
Taxonomy
TopicsPrivacy-Preserving Technologies in Data · Adversarial Robustness in Machine Learning · Advanced Neural Network Applications
