In this lecture, we introduce the diffusion models from the variational perspective. We will present the ideas of autoencoders (AEs), variational autoencoders (VAEs), denoising diffusion probabilistic models (DDPMs), discrete variational autoencoders (DVAEs), and discrete denoising diffusion probabilistic models (D3PMs).

1

In this lecture, we introduce the transformer architecture, the most widely used model in modern NLP.

2

In this lecture, we introduce the conjugate gradient method (CGM) for solving the system of linear equations.

3

My post is built based on Gregory Gundersen's blog-theme.

4