diff-ae.github.io - Diffusion Autoencoders: Toward a Meaningful and Decodable Representation

home > domains > diff-ae.github.io

Example domain paragraphs

Diffusion probabilistic models (DPMs) have achieved remarkable quality in image generation that rivals GANs'. But unlike GANs, DPMs use a set of latent variables that lack semantic meaning and cannot serve as a useful representation for other tasks. This paper explores the possibility of using DPMs for representation learning and seeks to extract a meaningful and decodable representation of an input image via autoencoding. Our key idea is to use a learnable encoder for discovering the high-level semantics,

Links to diff-ae.github.io (2)

• supasorn.com Supasorn Suwajanakorn
• diffusion-face-relighting.github.io Diffusion Face Relighting