Abstract: The transferability of adversarial examples allows for the attack on unknown deep neural networks (DNNs), posing a serious threat to many applications and attracting great attention. In this ...
CrowdStrike's 2025 data shows attackers breach AI systems in 51 seconds. Field CISOs reveal how inference security platforms defend against prompt injection, model extraction, and 9 other runtime ...
Abstract: This paper addresses the attack detection problem for cyber-physical systems subject to false data injection attacks. A novel detection framework is developed for cyber-physical systems ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results