Lifted Relational Variational Inference

Jaesik Choi, Eyal Amir
Proceedings of the 28th Conference on Uncertainty in Artificial Intelligence, PMLR R10:194-204, 2012.

Abstract

Hybrid continuous-discrete models naturally represent many real-world applications in robotics, finance, and environmental engineering. Inference with large-scale models is challenging because relational structures deteriorate rapidly during inference with observations. The main contribution of this paper is an efficient relational variational inference algorithm that factors largescale probability models into simpler variational models, composed of mixtures of iid (Bernoulli) random variables. The algorithm takes probability relational models of largescale hybrid systems and converts them to a close-to-optimal variational models. Then, it efficiently calculates marginal probabilities on the variational models by using a latent (or lifted) variable elimination or a lifted stochastic sampling. This inference is unique because it maintains the relational structure upon individual observations and during inference steps.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR10-choi12a, title = {Lifted Relational Variational Inference}, author = {Choi, Jaesik and Amir, Eyal}, booktitle = {Proceedings of the 28th Conference on Uncertainty in Artificial Intelligence}, pages = {194--204}, year = {2012}, editor = {de Freitas, Nando and Murphy, Kevin}, volume = {R10}, series = {Proceedings of Machine Learning Research}, month = {14--18 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r10/main/assets/choi12a/choi12a.pdf}, url = {https://proceedings.mlr.press/r10/choi12a.html}, abstract = {Hybrid continuous-discrete models naturally represent many real-world applications in robotics, finance, and environmental engineering. Inference with large-scale models is challenging because relational structures deteriorate rapidly during inference with observations. The main contribution of this paper is an efficient relational variational inference algorithm that factors largescale probability models into simpler variational models, composed of mixtures of iid (Bernoulli) random variables. The algorithm takes probability relational models of largescale hybrid systems and converts them to a close-to-optimal variational models. Then, it efficiently calculates marginal probabilities on the variational models by using a latent (or lifted) variable elimination or a lifted stochastic sampling. This inference is unique because it maintains the relational structure upon individual observations and during inference steps.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Lifted Relational Variational Inference %A Jaesik Choi %A Eyal Amir %B Proceedings of the 28th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2012 %E Nando de Freitas %E Kevin Murphy %F pmlr-vR10-choi12a %I PMLR %P 194--204 %U https://proceedings.mlr.press/r10/choi12a.html %V R10 %X Hybrid continuous-discrete models naturally represent many real-world applications in robotics, finance, and environmental engineering. Inference with large-scale models is challenging because relational structures deteriorate rapidly during inference with observations. The main contribution of this paper is an efficient relational variational inference algorithm that factors largescale probability models into simpler variational models, composed of mixtures of iid (Bernoulli) random variables. The algorithm takes probability relational models of largescale hybrid systems and converts them to a close-to-optimal variational models. Then, it efficiently calculates marginal probabilities on the variational models by using a latent (or lifted) variable elimination or a lifted stochastic sampling. This inference is unique because it maintains the relational structure upon individual observations and during inference steps. %Z Reissued by PMLR on 04 October 2026.
APA
Choi, J. & Amir, E.. (2012). Lifted Relational Variational Inference. Proceedings of the 28th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R10:194-204 Available from https://proceedings.mlr.press/r10/choi12a.html. Reissued by PMLR on 04 October 2026.

Related Material