SoK: Unintended Interactions among Machine Learning Defenses and Risks

arXiv:2312.04542v2 Announce Type: replace-cross
Abstract: Machine learning (ML) models cannot neglect risks to security, privacy, and fairness. Several defenses have been proposed to mitigate such risks. When a defense is effective in mitigating one risk, it may correspond to increased or decreased susceptibility to other risks. Existing research lacks an effective framework to recognize and explain these unintended interactions. We present such a framework, based on the conjecture that overfitting and memorization underlie unintended interactions. We survey existing literature on …

abstract arxiv cs.cr cs.lg defense fairness interactions machine machine learning privacy research risk risks security type