<ul><li>Diffusion-based generative models, such as Denoising Diffusion Probabilistic Models (DDPMs), have achieved remarkable success in image generation.</li><li>A new interpretable diffusion model called Patronus is introduced, which integrates a prototypical network into DDPMs.</li><li>Patronus enhances interpretability by showing the learned prototypes and how they influence the generation process.</li><li>The model supports downstream tasks like image manipulation and can reveal shortcut learning in the generation process.</li></ul>

Patronus: Bringing Transparency to Diffusion Models with Prototypes

Discover more