Breaking Fair Binary Classification with Optimal Flipping Attacks
Minimizing risk with fairness constraints is one of the popular approaches to learning a fair classifier. Recent works showed that this approach yields an unfair classifier if the training set is corrupted. In this work, we study the minimum amount of data corruption required for a successful flippi...
Saved in:
Published in | 2022 IEEE International Symposium on Information Theory (ISIT) pp. 1453 - 1458 |
---|---|
Main Authors | , , |
Format | Conference Proceeding |
Language | English |
Published |
IEEE
26.06.2022
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | Minimizing risk with fairness constraints is one of the popular approaches to learning a fair classifier. Recent works showed that this approach yields an unfair classifier if the training set is corrupted. In this work, we study the minimum amount of data corruption required for a successful flipping attack. First, we find lower/upper bounds on this quantity and show that these bounds are tight when the target model is the unique unconstrained risk minimizer. Second, we propose a computationally efficient data poisoning attack algorithm that can compromise the performance of fair learning algorithms. |
---|---|
ISSN: | 2157-8117 |
DOI: | 10.1109/ISIT50566.2022.9834475 |