Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics

Joanna Marks, Tim Y. J. Wang, Omer Deniz Akyildiz
Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, PMLR 337:4299-4322, 2026.

Abstract

We develop interacting particle algorithms for learning latent variable models with energy-based priors. To do so, we leverage recent developments in particle-based methods for solving maximum marginal likelihood estimation (MMLE) problems. Specifically, we provide a continuous-time framework for learning latent energy-based models, by defining stochastic differential equations (SDEs) that provably solve the MMLE problem. We obtain a practical algorithm as a discretisation of these SDEs and provide theoretical guarantees for the convergence of the proposed algorithm. Finally, we demonstrate the empirical effectiveness of our method on synthetic and image datasets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v337-marks26a, title = {Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics}, author = {Marks, Joanna and Wang, Tim Y. J. and Akyildiz, Omer Deniz}, booktitle = {Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence}, pages = {4299--4322}, year = {2026}, editor = {Perković, Emilija and Malinsky, Daniel}, volume = {337}, series = {Proceedings of Machine Learning Research}, month = {17--21 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v337/main/assets/marks26a/marks26a.pdf}, url = {https://proceedings.mlr.press/v337/marks26a.html}, abstract = {We develop interacting particle algorithms for learning latent variable models with energy-based priors. To do so, we leverage recent developments in particle-based methods for solving maximum marginal likelihood estimation (MMLE) problems. Specifically, we provide a continuous-time framework for learning latent energy-based models, by defining stochastic differential equations (SDEs) that provably solve the MMLE problem. We obtain a practical algorithm as a discretisation of these SDEs and provide theoretical guarantees for the convergence of the proposed algorithm. Finally, we demonstrate the empirical effectiveness of our method on synthetic and image datasets.} }
Endnote
%0 Conference Paper %T Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics %A Joanna Marks %A Tim Y. J. Wang %A Omer Deniz Akyildiz %B Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Emilija Perković %E Daniel Malinsky %F pmlr-v337-marks26a %I PMLR %P 4299--4322 %U https://proceedings.mlr.press/v337/marks26a.html %V 337 %X We develop interacting particle algorithms for learning latent variable models with energy-based priors. To do so, we leverage recent developments in particle-based methods for solving maximum marginal likelihood estimation (MMLE) problems. Specifically, we provide a continuous-time framework for learning latent energy-based models, by defining stochastic differential equations (SDEs) that provably solve the MMLE problem. We obtain a practical algorithm as a discretisation of these SDEs and provide theoretical guarantees for the convergence of the proposed algorithm. Finally, we demonstrate the empirical effectiveness of our method on synthetic and image datasets.
APA
Marks, J., Wang, T.Y.J. & Akyildiz, O.D.. (2026). Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics. Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 337:4299-4322 Available from https://proceedings.mlr.press/v337/marks26a.html.

Related Material