Interacting Particle Markov Chain Monte Carlo

Tom Rainforth, Christian Naesseth, Fredrik Lindsten, Brooks Paige, Jan-Willem Vandemeent, Arnaud Doucet, Frank Wood
Proceedings of The 33rd International Conference on Machine Learning, PMLR 48:2616-2625, 2016.

Abstract

We introduce interacting particle Markov chain Monte Carlo (iPMCMC), a PMCMC method based on an interacting pool of standard and conditional sequential Monte Carlo samplers. Like related methods, iPMCMC is a Markov chain Monte Carlo sampler on an extended space. We present empirical results that show significant improvements in mixing rates relative to both non-interacting PMCMC samplers and a single PMCMC sampler with an equivalent memory and computational budget. An additional advantage of the iPMCMC method is that it is suitable for distributed and multi-core architectures.

Cite this Paper


BibTeX
@InProceedings{pmlr-v48-rainforth16, title = {Interacting Particle Markov Chain Monte Carlo}, author = {Rainforth, Tom and Naesseth, Christian and Lindsten, Fredrik and Paige, Brooks and Vandemeent, Jan-Willem and Doucet, Arnaud and Wood, Frank}, booktitle = {Proceedings of The 33rd International Conference on Machine Learning}, pages = {2616--2625}, year = {2016}, editor = {Balcan, Maria Florina and Weinberger, Kilian Q.}, volume = {48}, series = {Proceedings of Machine Learning Research}, address = {New York, New York, USA}, month = {20--22 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v48/rainforth16.pdf}, url = {https://proceedings.mlr.press/v48/rainforth16.html}, abstract = {We introduce interacting particle Markov chain Monte Carlo (iPMCMC), a PMCMC method based on an interacting pool of standard and conditional sequential Monte Carlo samplers. Like related methods, iPMCMC is a Markov chain Monte Carlo sampler on an extended space. We present empirical results that show significant improvements in mixing rates relative to both non-interacting PMCMC samplers and a single PMCMC sampler with an equivalent memory and computational budget. An additional advantage of the iPMCMC method is that it is suitable for distributed and multi-core architectures.} }
Endnote
%0 Conference Paper %T Interacting Particle Markov Chain Monte Carlo %A Tom Rainforth %A Christian Naesseth %A Fredrik Lindsten %A Brooks Paige %A Jan-Willem Vandemeent %A Arnaud Doucet %A Frank Wood %B Proceedings of The 33rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2016 %E Maria Florina Balcan %E Kilian Q. Weinberger %F pmlr-v48-rainforth16 %I PMLR %P 2616--2625 %U https://proceedings.mlr.press/v48/rainforth16.html %V 48 %X We introduce interacting particle Markov chain Monte Carlo (iPMCMC), a PMCMC method based on an interacting pool of standard and conditional sequential Monte Carlo samplers. Like related methods, iPMCMC is a Markov chain Monte Carlo sampler on an extended space. We present empirical results that show significant improvements in mixing rates relative to both non-interacting PMCMC samplers and a single PMCMC sampler with an equivalent memory and computational budget. An additional advantage of the iPMCMC method is that it is suitable for distributed and multi-core architectures.
RIS
TY - CPAPER TI - Interacting Particle Markov Chain Monte Carlo AU - Tom Rainforth AU - Christian Naesseth AU - Fredrik Lindsten AU - Brooks Paige AU - Jan-Willem Vandemeent AU - Arnaud Doucet AU - Frank Wood BT - Proceedings of The 33rd International Conference on Machine Learning DA - 2016/06/11 ED - Maria Florina Balcan ED - Kilian Q. Weinberger ID - pmlr-v48-rainforth16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 48 SP - 2616 EP - 2625 L1 - http://proceedings.mlr.press/v48/rainforth16.pdf UR - https://proceedings.mlr.press/v48/rainforth16.html AB - We introduce interacting particle Markov chain Monte Carlo (iPMCMC), a PMCMC method based on an interacting pool of standard and conditional sequential Monte Carlo samplers. Like related methods, iPMCMC is a Markov chain Monte Carlo sampler on an extended space. We present empirical results that show significant improvements in mixing rates relative to both non-interacting PMCMC samplers and a single PMCMC sampler with an equivalent memory and computational budget. An additional advantage of the iPMCMC method is that it is suitable for distributed and multi-core architectures. ER -
APA
Rainforth, T., Naesseth, C., Lindsten, F., Paige, B., Vandemeent, J., Doucet, A. & Wood, F.. (2016). Interacting Particle Markov Chain Monte Carlo. Proceedings of The 33rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 48:2616-2625 Available from https://proceedings.mlr.press/v48/rainforth16.html.

Related Material