Neurosymbolic AI. 3 credits, 3 contact hours

Catalog description

Prerequisites: CS 370 and CS 375 with a grade of C or better. This course introduces principles and methods for integrating neural networks with symbolic reasoning to build AI systems that are more robust, explainable, and adaptable. It covers representations and reasoning in logical, probabilistic, and causal frameworks, along with differentiable approaches to reasoning over discrete structures and performing structured inference and optimization. The course then examines how learning-based models can be combined with these forms of prior knowledge and structured world representations in modern AI systems. CS 477 is recommended but not required.