Saved in:
Bibliographic Details
Main Authors: Wang, Tim Y. J., Kuntz, Juan, Akyildiz, O. Deniz
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.12412
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908918098165760
author Wang, Tim Y. J.
Kuntz, Juan
Akyildiz, O. Deniz
author_facet Wang, Tim Y. J.
Kuntz, Juan
Akyildiz, O. Deniz
contents We introduce a novel particle-based algorithm for end-to-end training of latent diffusion models. We reformulate the training task as minimizing a free energy functional and obtain a gradient flow that does so. By approximating the latter with a system of interacting particles, we obtain the algorithm, which we underpin theoretically by providing error guarantees. The novel algorithm compares favorably in experiments with previous particle-based methods and variational inference analogues.
format Preprint
id arxiv_https___arxiv_org_abs_2505_12412
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Training Latent Diffusion Models with Interacting Particle Algorithms
Wang, Tim Y. J.
Kuntz, Juan
Akyildiz, O. Deniz
Machine Learning
We introduce a novel particle-based algorithm for end-to-end training of latent diffusion models. We reformulate the training task as minimizing a free energy functional and obtain a gradient flow that does so. By approximating the latter with a system of interacting particles, we obtain the algorithm, which we underpin theoretically by providing error guarantees. The novel algorithm compares favorably in experiments with previous particle-based methods and variational inference analogues.
title Training Latent Diffusion Models with Interacting Particle Algorithms
topic Machine Learning
url https://arxiv.org/abs/2505.12412