Research on Detection Methods for Adversarial Capability of Artificial Intelligence Software
Yanjiao Ma, Yanhong Chen, Yaofeng Nie, Manshuang She, Jielin Cheng, Tinghua Zhang · 2024
Artificial intelligence software faces more diverse network security attacks, and its Adversarial ability is crucial. This paper studies the Adversarial ability detection method of AI software and its application in the field of medical software, and proposes an AI software attack test specification scheme for the problem of difficult to measure the strength of the attack test, which provides the testers with the reference ideas for the management of the attack test.