Search Results

Machine learning (ML) are data-centric artificial intelligence (AI) methods that became popular as an approach to produce value from data. ML methods produce specific decisions that are narrow and opaque. The more mature field of interpretable ML has provided principles to the recent field of explainable AI (XAI) to produce explanations and make ML methods more transparent and to support user needs with explanations. This course will educate students on the motivations, methods, and the value of XAI. The emphasis is on approaches that help humans comprehend various facets of a ML method such as its decisions, its strengths and weaknesses, and its evaluation. The XAI methods studied are grouped in categories of perturbations, masks, gradient, propagation, intrinsic, twins, proxy, and instance attributions.

College/Department: College of Computing and Informatics
Repeat Status: Not repeatable for credit
Prerequisites: INFO 629 [Min Grade: C] or CS 510 [Min Grade: C] or CS 613 [Min Grade: C] or CS 615 [Min Grade: C]