Confidence Matters: Inspecting Backdoors in Deep Neural Networks via Distribution Transfer
원문 초록
Backdoor attacks have been shown to be a serious security threat against deep learning models, and detecting whether a given model has been backdoored becomes a crucial task. Existing defenses are mainly built upon the observation that the backdoor trigger is usually of small size or affects the activation of only a few neurons. However, the above observations are violated in many cases especially for advanced backdoor attacks, hindering the performance and applicability of the existing defenses. In this paper, we propose a backdoor defense DTInspector built upon a new observation. That is, an effective backdoor attack usually requires high prediction confidence on the poisoned training samples, so as to ensure that the trained model exhibits the targeted behavior with a high probability. Based on this observation, DTInspector first learns a patch that could change the predictions of most high-confidence data, and then decides the existence of backdoor by checking the ratio of prediction changes after applying the learned patch on the low-confidence data. Extensive evaluations on five backdoor attacks, four datasets, and three advanced attacking types demonstrate the effectiveness of the proposed defense.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Gungnir: Exploiting Stylistic Features in Images for Backdoor Attacks on Diffusion Models
"No Matter What You Do": Purifying GNN Models via Backdoor Unlearning
Diff-Cleanse: Identifying and Mitigating Backdoor Attacks in Diffusion Models
Towards Unified Robustness Against Both Backdoor and Adversarial Attacks
Mitigating Backdoor Attack by Injecting Proactive Defensive Backdoor
Watch Out for Your Agents! Investigating Backdoor Threats to LLM-Based Agents
토론
이 논문과 연결된 포럼 글이 아직 없습니다.