18-789 Deep Generative Modeling

12 units, Carnegie Mellon University

This course will explore the basics of deep generative modeling. It will cover state-of-the-art models for image and text/sequence generation, including generative adversarial networks (GANs), variational autoencoders (VAEs), diffusion models, recurrent neural networks (RNNs), and transformers. Students will learn the relative advantages and disadvantages of different model choices, as well as the fundamental design choices that went into each idea. Students will get a chance to explore and present cutting-edge research and will also implement and experiment with generative models through a course project.

Prerequisites: 18-461 (min B) or 18-661 (min B)

Offered: Spring 2027, Spring 2026

Courseweave needs JavaScript to show sections, workloads and planning tools.