Transport Reversible Jump Proposals

Laurence Davies, Robert Salomone, Matthew Sutton, Chris Drovandi
Proceedings of The 26th International Conference on Artificial Intelligence and Statistics, PMLR 206:6839-6852, 2023.

Abstract

Reversible jump Markov chain Monte Carlo (RJMCMC) proposals that achieve reasonable acceptance rates and mixing are notoriously difficult to design in most applications. Inspired by recent advances in deep neural network-based normalizing flows and density estimation, we demonstrate an approach to enhance the efficiency of RJMCMC sampling by performing transdimensional jumps involving reference distributions. In contrast to other RJMCMC proposals, the proposed method is the first to apply a non-linear transport-based approach to construct efficient proposals between models with complicated dependency structures. It is shown that, in the setting where exact transports are used, our RJMCMC proposals have the desirable property that the acceptance probability depends only on the model probabilities. Numerical experiments demonstrate the efficacy of the approach.

Cite this Paper


BibTeX
@InProceedings{pmlr-v206-davies23a, title = {Transport Reversible Jump Proposals}, author = {Davies, Laurence and Salomone, Robert and Sutton, Matthew and Drovandi, Chris}, booktitle = {Proceedings of The 26th International Conference on Artificial Intelligence and Statistics}, pages = {6839--6852}, year = {2023}, editor = {Ruiz, Francisco and Dy, Jennifer and van de Meent, Jan-Willem}, volume = {206}, series = {Proceedings of Machine Learning Research}, month = {25--27 Apr}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v206/davies23a/davies23a.pdf}, url = {https://proceedings.mlr.press/v206/davies23a.html}, abstract = {Reversible jump Markov chain Monte Carlo (RJMCMC) proposals that achieve reasonable acceptance rates and mixing are notoriously difficult to design in most applications. Inspired by recent advances in deep neural network-based normalizing flows and density estimation, we demonstrate an approach to enhance the efficiency of RJMCMC sampling by performing transdimensional jumps involving reference distributions. In contrast to other RJMCMC proposals, the proposed method is the first to apply a non-linear transport-based approach to construct efficient proposals between models with complicated dependency structures. It is shown that, in the setting where exact transports are used, our RJMCMC proposals have the desirable property that the acceptance probability depends only on the model probabilities. Numerical experiments demonstrate the efficacy of the approach.} }
Endnote
%0 Conference Paper %T Transport Reversible Jump Proposals %A Laurence Davies %A Robert Salomone %A Matthew Sutton %A Chris Drovandi %B Proceedings of The 26th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2023 %E Francisco Ruiz %E Jennifer Dy %E Jan-Willem van de Meent %F pmlr-v206-davies23a %I PMLR %P 6839--6852 %U https://proceedings.mlr.press/v206/davies23a.html %V 206 %X Reversible jump Markov chain Monte Carlo (RJMCMC) proposals that achieve reasonable acceptance rates and mixing are notoriously difficult to design in most applications. Inspired by recent advances in deep neural network-based normalizing flows and density estimation, we demonstrate an approach to enhance the efficiency of RJMCMC sampling by performing transdimensional jumps involving reference distributions. In contrast to other RJMCMC proposals, the proposed method is the first to apply a non-linear transport-based approach to construct efficient proposals between models with complicated dependency structures. It is shown that, in the setting where exact transports are used, our RJMCMC proposals have the desirable property that the acceptance probability depends only on the model probabilities. Numerical experiments demonstrate the efficacy of the approach.
APA
Davies, L., Salomone, R., Sutton, M. & Drovandi, C.. (2023). Transport Reversible Jump Proposals. Proceedings of The 26th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 206:6839-6852 Available from https://proceedings.mlr.press/v206/davies23a.html.

Related Material