On Adversarial Training with Incorrect Labels

Benjamin Zi Hao Zhao, Junda Lu, Xiaowei Zhou, Dinusha Vatsalan, Muhammad Ikram, Mohamed Ali Kâafar

Abstract: In this work, we study adversarial training in the presence of incorrectly labeled data. Specifically, the predictive performance of an adversarially trained Machine Learning (ML) model trained on clean data and when the labels of training data and adversarial examples contain erroneous labels. Such erroneous labels may arise organically from a flawed labeling process or maliciously akin to a poisoning attacker.