Structured Generations: Using Hierarchical Clusters to guide Diffusion Models

TL;DR

This paper presents Diffuse-TreeVAE, a hierarchical generative model combining latent tree structures with diffusion models to produce high-quality, cluster-representative images.

Contribution

It introduces a novel hierarchical clustering-guided diffusion framework that enhances image quality and cluster fidelity over previous VAE-based methods.

Findings

Improved image clarity and diversity.

Generated samples accurately reflect data clusters.

Addresses limitations of prior VAE-based generative models.

Abstract

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.