Qianru Wu

1 month ago | mdpi.com | Ran Zhang |Qianru Wu |Yifan Wang

1. IntroductionDeep learning technology is now advancing rapidly in many fields. With the continuous promotion and popularization of deep learning technology, deep learning models have become targets of attacks, and their security has gradually become an obstacle to the future development and application of deep learning technology [1,2,3]. Common security threats in the field of deep learning mainly include data poisoning, backdoor attacks, adversarial example attacks, etc.