Course on Diffusion Models for Generative AI

Sanjay Shakkottai delivers lectures on the mathematical foundations of Diffusion Generative AI models. The lecture videos are posted on Tuesdays and Thursdays through the Fall 2025 semester. 

Full playlist of lectures is here.
Lecture notes for this series are available here. 


Lecture 01: Sampling Fundamentals (Rejection., Metropolis-Hastings and Gibbs)

Lecture 02: Langevin Dynamics and Introduction to SDEs

Lecture 03: Markov Processes

Lecture 04: Markov Processes, Divergence, and the Poincare Inequality

Lecture 05: Poincare Inequality and the Carre du Champ operator

Lecture 06: Bakry Emery Criterion 

Lecture 07: Log Sobolev Inequality / Score and MMSE

Lecture 08: Variational Auto Encoders and GANs

Lecture 09: DDPM and Tweedie’s Formula (continuous time)

Lecture 10: DDIM + Discrete-time Denoising through ELBO

Lecture 11: Discrete-time Denoising through ELBO, Intro to Ito Calculus

Lecture 12: Girsanov’s Theorem and Intro to discretization