COSC 89.31 Deep Learning Generalization and Robustness
This course will be an extended version of COSC 78/278 Deep Learning. It is mostly project-based, and it aims to bridge the gap between machine learning course materials and recent developments in machine learning research. The course begins by covering the basics of model training and inference. From there, the course proceeds to discuss various concepts of generalization, different types of robustness issues associated with generalization (e.g., adversarial robustness, out-of-distribution robustness, model poisoning, etc.), and the connections between (robust) generalization to the design of multiple regularization and normalization strategies.