Training Latent Diffusion Models with Interacting Particle Algorithms

2025/05/18 by Wang, Tim Y. J., Kuntz, Juan, Akyildiz, O. Deniz · 1 citation
#FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML)

paper · doi:10.48550/arxiv.2505.12412

Abstract

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.

Cited by

• Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics

Related

• Particle algorithms for maximum likelihood training of latent variable models
• Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics
• Diffusion map particle systems for generative modeling
• Momentum Particle Maximum Likelihood
• Particle Dynamics for Latent-Variable Energy-Based Models