MIT 6.S184: Flow Matching and Diffusion Models - Lecture 1 - Generative AI with SDEs

MIT 6.S184 An Introduction to Flow and Diffusion Models, IAP 2025
Instructors: Peter Holderrieth and Ezra Erives
YouTube playlist:    • MIT 6.S184: An Introduction to Flow M...  
View the complete course materials, including homework: https://diffusion.csail.mit.edu
Lecture notes: https://diffusion.csail.mit.edu/docs/...

More courses at https://soul.mit.edu.
Support SOUL at https://mitsoul.org/donate/.

(We have posted this course both on the instructor's YouTube channel, and also on this channel. The videos are identical.)

Course description:
Diffusion and flow-based models have become the state of the art for generative AI across a wide range of data modalities, including images, videos, shapes, molecules, music, and more! This course aims to build up the mathematical framework underlying these models from first principles. At the end of the class, students will have built a toy image diffusion model from scratch, and along the way, will have gained hands-on experience with the mathematical toolbox of stochastic differential equations that is useful in many other fields. This course is ideal for students who want to develop a principled understanding of the theory and practice of generative AI.

A note about the recording quality:
These recordings were made using a retrofit automatic recording system (https://canvas.mit.edu/courses/3156/p...) that is relatively inexpensive. Ideally, we would've been able to record these lectures at much higher video and audio quality. If you want to support such efforts, consider donating.

We encourage constructive comments and discussion on our YouTube and other social media channels. Personal attacks, hate speech, trolling, and inappropriate comments are not allowed and may be removed. We follow the OCW's policy on comments available here: https://ocw.mit.edu/comments.