<aside> 📣 Announcements: Please submit your course evaluation at: https://ucsb.bluera.com/ucsb (Due Mar 15 ‼️)
</aside>
Generative models are disrupting the world in multiple domains, including images, text, and speech. In this course, we will cover the probabilistic foundations of deep generative models. We will learn traditional and state-of-the-art models including autoregressive models, variational autoencoders, generative adversarial networks, energy-based models, and diffusion models. The course will also discuss interesting application areas such as deep generative models for data compression and scientific discovery.
This course and lecture materials are based off of Stanford’s CS 236.
Prerequisites: Basic knowledge of machine learning, including adequate knowledge of probability, statistics, linear algebra, and optimization, and proficiency with Python. This course will move fast assuming that you have already taken one or more machine learning classes.