Rejection-Sampling-Based Ancestor Sampling for Particle Gibbs

Roland Hostettler; Simo Särkkä

doi:10.1109/MLSP.2019.8918852

Rejection-Sampling-Based Ancestor Sampling for Particle Gibbs

Roland Hostettler
, Simo Särkkä

Uppsala University

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

Abstract

Particle Gibbs with ancestor sampling is an efficient and statistically principled algorithm for learning of dynamic systems. However, the ancestor sampling step used to improve mixing of the Markov chain requires the possibly expensive calculation of a set of ancestor weights for the complete particle system. In this paper, we propose a rejection-sampling-based algorithm for ancestor sampling in particle Gibbs that mitigates this problem. Additionally, performance guarantees and a fallback strategy to prevent suffering from high rejection rates are discussed. The performance of the method is illustrated in two numerical examples.

Original language	English
Title of host publication	Proceedings of the 29th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2019
Publisher	IEEE
Number of pages	6
ISBN (Electronic)	9781728108247
DOIs	https://doi.org/10.1109/MLSP.2019.8918852
Publication status	Published - 1 Oct 2019
MoE publication type	A4 Conference publication
Event	IEEE International Workshop on Machine Learning for Signal Processing - Pittsburgh, United States Duration: 13 Oct 2019 → 16 Oct 2019 Conference number: 29

Publication series

Name	IEEE International Workshop on Machine Learning for Signal Processing
Publisher	IEEE
ISSN (Print)	2161-0363
ISSN (Electronic)	2161-0371

Workshop

Workshop	IEEE International Workshop on Machine Learning for Signal Processing
Abbreviated title	MLSP
Country/Territory	United States
City	Pittsburgh
Period	13/10/2019 → 16/10/2019

Funding

Financial support by the Academy of Finland is gratefully acknowledged.

Keywords

Particle filters
Particle Markov chain Monte Carlo
Statistical learning

Access to Document

10.1109/MLSP.2019.8918852

Cite this

@inproceedings{a8cc8558bab34564995776cee1e4a510,

title = "Rejection-Sampling-Based Ancestor Sampling for Particle Gibbs",

abstract = "Particle Gibbs with ancestor sampling is an efficient and statistically principled algorithm for learning of dynamic systems. However, the ancestor sampling step used to improve mixing of the Markov chain requires the possibly expensive calculation of a set of ancestor weights for the complete particle system. In this paper, we propose a rejection-sampling-based algorithm for ancestor sampling in particle Gibbs that mitigates this problem. Additionally, performance guarantees and a fallback strategy to prevent suffering from high rejection rates are discussed. The performance of the method is illustrated in two numerical examples.",

keywords = "Particle filters, Particle Markov chain Monte Carlo, Statistical learning",

author = "Roland Hostettler and Simo S{\"a}rkk{\"a}",

year = "2019",

month = oct,

day = "1",

doi = "10.1109/MLSP.2019.8918852",

language = "English",

series = "IEEE International Workshop on Machine Learning for Signal Processing",

publisher = "IEEE",

booktitle = "Proceedings of the 29th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2019",

address = "United States",

note = "IEEE International Workshop on Machine Learning for Signal Processing, MLSP ; Conference date: 13-10-2019 Through 16-10-2019",

}

Hostettler, R & Särkkä, S 2019, Rejection-Sampling-Based Ancestor Sampling for Particle Gibbs. in Proceedings of the 29th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2019., 8918852, IEEE International Workshop on Machine Learning for Signal Processing, IEEE, IEEE International Workshop on Machine Learning for Signal Processing, Pittsburgh, Pennsylvania, United States, 13/10/2019. https://doi.org/10.1109/MLSP.2019.8918852

Rejection-Sampling-Based Ancestor Sampling for Particle Gibbs. / Hostettler, Roland; Särkkä, Simo.
Proceedings of the 29th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2019. IEEE, 2019. 8918852 (IEEE International Workshop on Machine Learning for Signal Processing).

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

TY - GEN

T1 - Rejection-Sampling-Based Ancestor Sampling for Particle Gibbs

AU - Hostettler, Roland

AU - Särkkä, Simo

N1 - Conference code: 29

PY - 2019/10/1

Y1 - 2019/10/1

N2 - Particle Gibbs with ancestor sampling is an efficient and statistically principled algorithm for learning of dynamic systems. However, the ancestor sampling step used to improve mixing of the Markov chain requires the possibly expensive calculation of a set of ancestor weights for the complete particle system. In this paper, we propose a rejection-sampling-based algorithm for ancestor sampling in particle Gibbs that mitigates this problem. Additionally, performance guarantees and a fallback strategy to prevent suffering from high rejection rates are discussed. The performance of the method is illustrated in two numerical examples.

AB - Particle Gibbs with ancestor sampling is an efficient and statistically principled algorithm for learning of dynamic systems. However, the ancestor sampling step used to improve mixing of the Markov chain requires the possibly expensive calculation of a set of ancestor weights for the complete particle system. In this paper, we propose a rejection-sampling-based algorithm for ancestor sampling in particle Gibbs that mitigates this problem. Additionally, performance guarantees and a fallback strategy to prevent suffering from high rejection rates are discussed. The performance of the method is illustrated in two numerical examples.

KW - Particle filters

KW - Particle Markov chain Monte Carlo

KW - Statistical learning

UR - https://www.scopus.com/pages/publications/85077707888

U2 - 10.1109/MLSP.2019.8918852

DO - 10.1109/MLSP.2019.8918852

M3 - Conference article in proceedings

AN - SCOPUS:85077707888

T3 - IEEE International Workshop on Machine Learning for Signal Processing

BT - Proceedings of the 29th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2019

PB - IEEE

T2 - IEEE International Workshop on Machine Learning for Signal Processing

Y2 - 13 October 2019 through 16 October 2019

ER -

Rejection-Sampling-Based Ancestor Sampling for Particle Gibbs

Abstract

Publication series

Workshop

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this