Benign on Label, Malicious by Design: Clean-Label Dormant-to-Activated Backdoor via Machine Unlearning with Removable Camouflage

본문 미리보기

arXiv:2607.27936v1 Announce Type: new Abstract: Existing backdoor attacks often become effective immediately after backdoor implantation and may therefore be exposed before exploitation. Machine unlearning activated dormant backdoors mitigate such behavioral exposure by remaining inactive after training and becoming effective only after selected training records are unlearned. However, existing methods struggle to simultaneously achieve a low pre-unlearning attack success rate and strong post-u