This paper introduces Diffuse-TreeVAE, a deep generative model that integrates hierarchical clustering into the framework of Denoising Diffusion Probabilistic Models (DDPMs). The proposed approach generates new images by sampling from a root embedding of a learned latent tree VAE-based structure, it then propagates through hierarchical paths, and utilizes a second-stage DDPM to refine and generate distinct, high-quality images for each data cluster. The result is a model that not only improves image clarity but also ensures that the generated samples are representative of their respective clusters, addressing the limitations of previous VAE-based methods and advancing the state of clustering-based generative modeling.
@article{arxiv.2407.06124,
title = {Structured Generations: Using Hierarchical Clusters to guide Diffusion Models},
author = {Jorge da Silva Goncalves and Laura Manduchi and Moritz Vandenhirtz and Julia E. Vogt},
journal= {arXiv preprint arXiv:2407.06124},
year = {2024}
}