AI-infused Security: Robust Defense by Bridging Theory and Practice

While Artificial Intelligence (AI) has tremendous potential as a defense against real-world cybersecurity threats, understanding the capabilities and robustness of AI remains a fundamental challenge, especially in adversarial environments. In this talk, I address two interrelated problems that are essential to deployment of AI in security settings. (1) Designing robust and efficient machine learning algorithms with strong theoretical guarantees for large-scale, distributed, and noisy data. Specifically, I will present a boosting-based learning approach that solves an open problem in distributed learning. Based on this boosting algorithm and its insightful connection to game theory, I then propose a novel online framework that balances between risk and reward in adversarial scenarios. (2) Discovering real-world vulnerabilities of deep neural networks in and countermeasures to mitigate threats. I will present ShapeShifter, the first targeted physical adversarial attack that fools state-of-the-art object detectors, and SHIELD, a real-time defense that removes adversarial noise by stochastic data compression. Finally, I share my vision on making AI more robust under different threat models, and research directions on deploying AI in security-critical and high-stakes societal problems, such as cyber threat detection and fraud detection.