In this talk, I will present a detailed research topic about AI Cybersecurity. Our group proposed an advanced gradient-based approach for mitigation of adversarial attacks in Deep Neural Networks (DNN). The proposed approach adopted a random distortion transformation defense method called RDG (Random Distortion over Grids) and we combined it with non-linear defenses to thwart adversarial attacks. Extensive evaluation demonstrated the efficiency of this state-of-art defense approach. Co-sponsored by: Fairleigh Dickinson University Speaker(s): Dr. Meikang Qiu, Agenda: In this talk, I will present a detailed research topic about AI Cybersecurity. Our group proposed an advanced gradient-based approach for mitigation of adversarial attacks in Deep Neural Networks (DNN). The proposed approach adopted a random distortion transformation defense method called RDG (Random Distortion over Grids) and we combined it with non-linear defenses to thwart adversarial attacks. Extensive evaluation demonstrated the efficiency of this state-of-art defense approach. Room: M105, Bldg: Muscarelle Center, M105, , 1000 River Road , Teaneck , New Jersey, United States, 07666, Virtual: https://events.vtools.ieee.org/m/408037