Building Diffusion Model's theory from ground up
Keywords: diffusion; score; generative modelling
Blogpost Url: https://iclr-blogposts.github.io/2024/blog/diffusion-theory-from-scratch/
Abstract: Diffusion Models, a new generative model family, have taken the world by storm after the seminal paper by Ho et al. [2020]. While diffusion models are often described as a probabilistic Markov Chains, their underlying principle is based on the decade-old theory of Stochastic Differential Equations (SDE), as found out later by Song et al. [2021]. In this article, we will go back and revisit the 'fundamental ingredients' behind the SDE formulation and show how the idea can be 'shaped' to get to the modern form of Score-based Diffusion Models. We'll start from the very definition of the 'score', how it was used in the context of generative modeling, how we achieve the necessary theoretical guarantees and how the critical design choices were made to finally arrive at the more 'principled' framework of Score-based Diffusion. Throughout this article, we provide several intuitive illustrations for ease of understanding.
Ref Papers: https://openreview.net/pdf/ef0eadbe07115b0853e964f17aa09d811cd490f1.pdf
Id Of The Authors Of The Papers: ~Yang_Song1
Conflict Of Interest: No conflict.
Submission Number: 5
Loading