SampDetox : Black-box backdoor defense via perturbation-based sample detoxification

• Main Authors: YANG, Yanxin, JIA, Chentao, YAN, Dengke, HU, Ming, LI, Tianlin, XIE, Xiaofei, WEI, Xian, CHEN, Mingsong
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2024
• Subjects: Machine learning; Backdoor threats; Backdoor defense; Information Security
• Online Access: https://ink.library.smu.edu.sg/sis_research/9812; https://ink.library.smu.edu.sg/context/sis_research/article/10812/viewcontent/8771_SampDetox_Black_box_Backd.pdf
• Institution: Singapore Management University

The advancement of Machine Learning has enabled the widespread deployment of Machine Learning as a Service (MLaaS) applications. However, the untrustworthy nature of third-party ML services poses backdoor threats. Existing defenses in MLaaS are limited by their reliance on training samples or white-...