Un-fair trojan: Targeted backdoor attacks against model fairness

Author

Nicholas Furth, New Jersey Institute of TechnologyFollow

Document Type

Thesis

Date of Award

5-31-2022

Degree Name

Master of Science in Computer Engineering - (M.S.)

Department

Electrical and Computer Engineering

First Advisor

Abdallah Khreishah

Second Advisor

Cong Wang

Third Advisor

Hai Nhat Phan

Fourth Advisor

Qing Gary Liu

Abstract

Machine learning models have been shown to be vulnerable against various backdoor and data poisoning attacks that adversely affect model behavior. Additionally, these attacks have been shown to make unfair predictions with respect to certain protected features. In federated learning, multiple local models contribute to a single global model communicating only using local gradients, the issue of attacks become more prevalent and complex. Previously published works revolve around solving these issues both individually and jointly. However, there has been little study on the effects of attacks against model fairness. Demonstrated in this work, a flexible attack, which we call Un-Fair Trojan, that targets model fairness while remaining stealthy can have devastating effects against machine learning models.

Recommended Citation

Furth, Nicholas, "Un-fair trojan: Targeted backdoor attacks against model fairness" (2022). Theses. 1996.
https://digitalcommons.njit.edu/theses/1996