On Evaluating Neural Network Backdoor Defenses
This work addresses the need for reliable evaluation of backdoor defenses in deep neural networks, which is crucial for security applications, but it is incremental as it focuses on analysis and guidelines rather than proposing a new defense.
The paper tackled the problem of evaluating neural network backdoor defenses by identifying common pitfalls in existing methods and conducting a comprehensive side-by-side evaluation using a prepared database of attacks, resulting in general guidelines for developing more robust defenses.
Deep neural networks (DNNs) demonstrate superior performance in various fields, including scrutiny and security. However, recent studies have shown that DNNs are vulnerable to backdoor attacks. Several defenses were proposed in the past to defend DNNs against such backdoor attacks. In this work, we conduct a critical analysis and identify common pitfalls in these existing defenses, prepare a comprehensive database of backdoor attacks, conduct a side-by-side evaluation of existing defenses against this database. Finally, we layout some general guidelines to help researchers develop more robust defenses in the future and avoid common mistakes from the past.