Electrical & Computer Engin 690STC - Foundations/Generative Models

This course offers a rigorous exploration of the mathematical foundations, algorithmic frameworks, and cutting-edge developments in generative modeling. Students will examine both classical and modern approaches, starting with probabilistic models such as Gaussian Mixture Models and progressing through advanced deep learning architectures including Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and Deep Autoregressive Models. The course delves into state-of-the-art techniques such as Normalizing Flows, Energy-Based Models, Diffusion and Score-Based Models, and emerging flow-matching methods. Through theoretical analysis, coding assignments, and research paper discussions, students will gain the expertise needed to critically evaluate, implement, and innovate within the rapidly evolving field of generative modeling.