Sequential core-set Monte Carlo. Beronov, B., Weilbach, C., Wood, F., & Campbell, T. In de Campos, C. & Maathuis, M. H., editors, Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, volume 161, of Proceedings of Machine Learning Research, pages 2165–2175, 27–30 Jul, 2021. PMLR. ![Sequential core-set Monte Carlo [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper ![Sequential core-set Monte Carlo [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Presentation Poster abstract bibtex 7 downloads Sequential Monte Carlo (SMC) is a general-purpose methodology for recursive Bayesian inference, and is widely used in state space modeling and probabilistic programming. Its resample-move variant reduces the variance of posterior estimates by interleaving Markov chain Monte Carlo (MCMC) steps for particle “rejuvenation”; but this requires accessing all past observations and leads to linearly growing memory size and quadratic computation cost. Under the assumption of exchangeability, we introduce sequential core-set Monte Carlo (SCMC), which achieves constant space and linear time by rejuvenating based on sparse, weighted subsets of past data. In contrast to earlier approaches, which uniformly subsample or throw away observations, SCMC uses a novel online version of a state-of-the-art Bayesian core-set algorithm to incrementally construct a nonparametric, data- and model-dependent variational representation of the unnormalized target density. Experiments demonstrate significantly reduced approximation errors at negligible additional cost.
@InProceedings{BER-21,
title={Sequential core-set Monte Carlo},
author={Beronov, Boyan and Weilbach, Christian and Wood, Frank and Campbell, Trevor},
booktitle={Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence},
pages={2165--2175},
year={2021},
editor={de Campos, Cassio and Maathuis, Marloes H.},
volume={161},
series={Proceedings of Machine Learning Research},
month={27--30 Jul},
publisher={PMLR},
pdf={https://proceedings.mlr.press/v161/beronov21a/beronov21a.pdf},
url={https://proceedings.mlr.press/v161/beronov21a.html},
url_Presentation={https://github.com/plai-group/bibliography/raw/master/presentations_posters/UAI2021_BER_presentation.pdf},
url_Poster={https://github.com/plai-group/bibliography/raw/master/presentations_posters/UAI2021_BER_poster.pdf},
support={D3M},
abstract={Sequential Monte Carlo (SMC) is a general-purpose methodology for recursive Bayesian inference, and is widely used in state space modeling and probabilistic programming. Its resample-move variant reduces the variance of posterior estimates by interleaving Markov chain Monte Carlo (MCMC) steps for particle “rejuvenation”; but this requires accessing all past observations and leads to linearly growing memory size and quadratic computation cost. Under the assumption of exchangeability, we introduce sequential core-set Monte Carlo (SCMC), which achieves constant space and linear time by rejuvenating based on sparse, weighted subsets of past data. In contrast to earlier approaches, which uniformly subsample or throw away observations, SCMC uses a novel online version of a state-of-the-art Bayesian core-set algorithm to incrementally construct a nonparametric, data- and model-dependent variational representation of the unnormalized target density. Experiments demonstrate significantly reduced approximation errors at negligible additional cost.}
}
Downloads: 7
{"_id":"k45aGxW7RphnJ5KbY","bibbaseid":"beronov-weilbach-wood-campbell-sequentialcoresetmontecarlo-2021","author_short":["Beronov, B.","Weilbach, C.","Wood, F.","Campbell, T."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"Sequential core-set Monte Carlo","author":[{"propositions":[],"lastnames":["Beronov"],"firstnames":["Boyan"],"suffixes":[]},{"propositions":[],"lastnames":["Weilbach"],"firstnames":["Christian"],"suffixes":[]},{"propositions":[],"lastnames":["Wood"],"firstnames":["Frank"],"suffixes":[]},{"propositions":[],"lastnames":["Campbell"],"firstnames":["Trevor"],"suffixes":[]}],"booktitle":"Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence","pages":"2165–2175","year":"2021","editor":[{"propositions":["de"],"lastnames":["Campos"],"firstnames":["Cassio"],"suffixes":[]},{"propositions":[],"lastnames":["Maathuis"],"firstnames":["Marloes","H."],"suffixes":[]}],"volume":"161","series":"Proceedings of Machine Learning Research","month":"27–30 Jul","publisher":"PMLR","pdf":"https://proceedings.mlr.press/v161/beronov21a/beronov21a.pdf","url":"https://proceedings.mlr.press/v161/beronov21a.html","url_presentation":"https://github.com/plai-group/bibliography/raw/master/presentations_posters/UAI2021_BER_presentation.pdf","url_poster":"https://github.com/plai-group/bibliography/raw/master/presentations_posters/UAI2021_BER_poster.pdf","support":"D3M","abstract":"Sequential Monte Carlo (SMC) is a general-purpose methodology for recursive Bayesian inference, and is widely used in state space modeling and probabilistic programming. Its resample-move variant reduces the variance of posterior estimates by interleaving Markov chain Monte Carlo (MCMC) steps for particle “rejuvenation”; but this requires accessing all past observations and leads to linearly growing memory size and quadratic computation cost. Under the assumption of exchangeability, we introduce sequential core-set Monte Carlo (SCMC), which achieves constant space and linear time by rejuvenating based on sparse, weighted subsets of past data. In contrast to earlier approaches, which uniformly subsample or throw away observations, SCMC uses a novel online version of a state-of-the-art Bayesian core-set algorithm to incrementally construct a nonparametric, data- and model-dependent variational representation of the unnormalized target density. Experiments demonstrate significantly reduced approximation errors at negligible additional cost.","bibtex":"@InProceedings{BER-21,\n title={Sequential core-set Monte Carlo},\n author={Beronov, Boyan and Weilbach, Christian and Wood, Frank and Campbell, Trevor},\n booktitle={Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence},\n pages={2165--2175},\n year={2021},\n editor={de Campos, Cassio and Maathuis, Marloes H.},\n volume={161},\n series={Proceedings of Machine Learning Research},\n month={27--30 Jul},\n publisher={PMLR},\n pdf={https://proceedings.mlr.press/v161/beronov21a/beronov21a.pdf},\n url={https://proceedings.mlr.press/v161/beronov21a.html},\n url_Presentation={https://github.com/plai-group/bibliography/raw/master/presentations_posters/UAI2021_BER_presentation.pdf},\n url_Poster={https://github.com/plai-group/bibliography/raw/master/presentations_posters/UAI2021_BER_poster.pdf},\n support={D3M},\n abstract={Sequential Monte Carlo (SMC) is a general-purpose methodology for recursive Bayesian inference, and is widely used in state space modeling and probabilistic programming. Its resample-move variant reduces the variance of posterior estimates by interleaving Markov chain Monte Carlo (MCMC) steps for particle “rejuvenation”; but this requires accessing all past observations and leads to linearly growing memory size and quadratic computation cost. Under the assumption of exchangeability, we introduce sequential core-set Monte Carlo (SCMC), which achieves constant space and linear time by rejuvenating based on sparse, weighted subsets of past data. In contrast to earlier approaches, which uniformly subsample or throw away observations, SCMC uses a novel online version of a state-of-the-art Bayesian core-set algorithm to incrementally construct a nonparametric, data- and model-dependent variational representation of the unnormalized target density. Experiments demonstrate significantly reduced approximation errors at negligible additional cost.}\n}\n\n","author_short":["Beronov, B.","Weilbach, C.","Wood, F.","Campbell, T."],"editor_short":["de Campos, C.","Maathuis, M. H."],"key":"BER-21","id":"BER-21","bibbaseid":"beronov-weilbach-wood-campbell-sequentialcoresetmontecarlo-2021","role":"author","urls":{"Paper":"https://proceedings.mlr.press/v161/beronov21a.html"," presentation":"https://github.com/plai-group/bibliography/raw/master/presentations_posters/UAI2021_BER_presentation.pdf"," poster":"https://github.com/plai-group/bibliography/raw/master/presentations_posters/UAI2021_BER_poster.pdf"},"metadata":{"authorlinks":{}},"downloads":7},"bibtype":"inproceedings","biburl":"https://raw.githubusercontent.com/plai-group/bibliography/master/group_publications.bib","dataSources":["BKH7YtW7K7WNMA3cj","7avRLRrz2ifJGMKcD","wyN5DxtoT6AQuiXnm"],"keywords":[],"search_terms":["sequential","core","set","monte","carlo","beronov","weilbach","wood","campbell"],"title":"Sequential core-set Monte Carlo","year":2021,"downloads":7}