Elliptical Slice Sampling with Expectation Propagation

Francois Fagan Columbia University, Jalaj Bhandari Columbia University, John Cunningham Columbia University
Proceedings of the 32nd Conference on Uncertainty in Artificial Intelligence, PMLR R14:425-434, 2016.

Abstract

Markov Chain Monte Carlo techniques remain the gold standard for approximate Bayesian inference, but their practical issues – including onerous runtime and sensitivity to tuning parameters – often lead researchers to use faster but typically less accurate deterministic approximations. Here we couple the fast but biased deterministic approximation offered by expectation propagation with elliptical slice sampling, a state-of-the-art MCMC method. We extend our hybrid deterministic-MCMC method to include recycled samples and analytical slices, and we rigorously prove the validity of each enhancement. Taken together, we show that these advances provide an order of magnitude gain in efficiency beyond existing state-of-the-art sampling techniques in Bayesian classification and multivariate gaussian quadrature problems.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR14-university16l, title = {Elliptical Slice Sampling with Expectation Propagation}, author = {University, Francois Fagan Columbia and University, Jalaj Bhandari Columbia and University, John Cunningham Columbia}, booktitle = {Proceedings of the 32nd Conference on Uncertainty in Artificial Intelligence}, pages = {425--434}, year = {2016}, editor = {Ihler, Alexander and Janzing, Dominik}, volume = {R14}, series = {Proceedings of Machine Learning Research}, month = {25--29 Jun}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r14/main/assets/university16l/university16l.pdf}, url = {https://proceedings.mlr.press/r14/university16l.html}, abstract = {Markov Chain Monte Carlo techniques remain the gold standard for approximate Bayesian inference, but their practical issues – including onerous runtime and sensitivity to tuning parameters – often lead researchers to use faster but typically less accurate deterministic approximations. Here we couple the fast but biased deterministic approximation offered by expectation propagation with elliptical slice sampling, a state-of-the-art MCMC method. We extend our hybrid deterministic-MCMC method to include recycled samples and analytical slices, and we rigorously prove the validity of each enhancement. Taken together, we show that these advances provide an order of magnitude gain in efficiency beyond existing state-of-the-art sampling techniques in Bayesian classification and multivariate gaussian quadrature problems.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Elliptical Slice Sampling with Expectation Propagation %A Francois Fagan Columbia University %A Jalaj Bhandari Columbia University %A John Cunningham Columbia University %B Proceedings of the 32nd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2016 %E Alexander Ihler %E Dominik Janzing %F pmlr-vR14-university16l %I PMLR %P 425--434 %U https://proceedings.mlr.press/r14/university16l.html %V R14 %X Markov Chain Monte Carlo techniques remain the gold standard for approximate Bayesian inference, but their practical issues – including onerous runtime and sensitivity to tuning parameters – often lead researchers to use faster but typically less accurate deterministic approximations. Here we couple the fast but biased deterministic approximation offered by expectation propagation with elliptical slice sampling, a state-of-the-art MCMC method. We extend our hybrid deterministic-MCMC method to include recycled samples and analytical slices, and we rigorously prove the validity of each enhancement. Taken together, we show that these advances provide an order of magnitude gain in efficiency beyond existing state-of-the-art sampling techniques in Bayesian classification and multivariate gaussian quadrature problems. %Z Reissued by PMLR on 04 October 2026.
APA
University, F.F.C., University, J.B.C. & University, J.C.C.. (2016). Elliptical Slice Sampling with Expectation Propagation. Proceedings of the 32nd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R14:425-434 Available from https://proceedings.mlr.press/r14/university16l.html. Reissued by PMLR on 04 October 2026.

Related Material