Qun Song | Towards Resilient Autonomous Cyber-Physical Systems against Adversarial Examples

Deep learning is shown to be susceptible to adversarial examples, which are crafted inputs aiming to cause wrong classification outputs for deep models by adding minute perturbations on the clean inputs. Thus, deploying deep learning models on safety-critical cyber-physical systems without incorporating effective countermeasures against adversarial examples raises security concerns. This talk is about the studies on the threat and countermeasures for the adversarial example attack as an ongoing concern for the safety-critical autonomous cyber-physical systems. This talk will introduce the dynamic ensemble-based defenses designed under the strategy of moving target defense that effectively counteract the adaptive adversarial example adversary for embedded deep visual sensing. This talk will also present the systematic requirement investigation and credibility analysis of adversarial example attack against the power grid voltage stability assessment and effective countermeasure.