Foundation models and generative AI

This week we explore denoising diffusion models that generate data through iterative refinement from noise, covering their mathematical foundations, connections to score-based methods, acceleration techniques, and applications.

This week we will have a guest lecture by Karsten Kreis from NVIDIA as well as Ruiqi Gao from DeepMind. To accommodate their time zone, the lecture will exceptionally take place at 5 pm in room CM 1 3.

Important! Assignment 1 is released. Please find below the instructions and information required for the submission. Information on gNoto can be found on top of the Moodle page. Assignment 1 is due on October 8, 23:59.

This Week's TA: Johann Wenckstern