Literature Review of Backdoor Attacks

In this chapter, we first introduce three application areas of deep neural networks, including computer vision, natural language processing, and federated learning-based e-Health. Then, we introduce the security model, i.e., white-, gray-, and black-box attack assumptions. We also give some basic definitions and specific implementations of backdoor attacks in deep neural networks. In addition, we introduce performance metrics for backdoor attacks, which mainly include the success rate of the attack and the availability of the model. At last, we survey related works on backdoor attacks to provide a comprehensive overview of the current literature on the three application areas of deep neural networks mentioned above.