2023 Fellow

Ryan Feng is a Ph.D. candidate in the Department of Computer Science and Engineering. He is co-advised by Professor Atul Prakash and Professor Stella Yu. His research lies in the intersection of computer security, machine learning, and computer vision, working to improve the robustness of machine learning and autonomous systems to make them more secure and trustworthy. His recent work includes GRAPHITE, an automatic framework for generating physical-world image perturbations in a black-box setting. The GRAPHITE framework can automatically generate useful test cases and guide defense design to defend machine learning models in settings such as traffic sign classification against such physical-world perturbations. His current work focuses on leveraging physical and semantic constraints extracted from data-driven statistical properties to force machine learning models to rely on more robust features.