Scalable importance sampling estimation of Gaussian mixture posteriors in Bayesian networks

被引：8

作者：

Ramos-Lopez, Dario ^{[1
]}

Masegosa, Andres R. ^{[1
]}

Salmeron, Antonio ^{[1
]}

Rumi, Rafael ^{[1
]}

Langseth, Helge ^{[3
]}

Nielsen, Thomas D. ^{[1
,2
]}

Madsen, Anders L. ^{[2
,4
]}

机构：

[1] Univ Almeria, Dept Math, Almeria, Spain

[2] Aalborg Univ, Dept Comp Sci, Aalborg, Denmark

[3] Norwegian Univ Sci & Technol, Dept Comp & Informat Sci, Trondheim, Norway

[4] HUGIN EXPERT AS, Aalborg, Denmark

来源：

INTERNATIONAL JOURNAL OF APPROXIMATE REASONING | 2018年 / 100卷

关键词：

Importance sampling; Bayesian networks; Conditional linear Gaussian models; Scalable inference; Gaussian mixtures; INFERENCE; PROPAGATION; COMPUTATION; ALGORITHM;

D O I：

10.1016/j.ijar.2018.06.004

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In this paper we propose a scalable importance sampling algorithm for computing Gaussian mixture posteriors in conditional linear Gaussian Bayesian networks. Our contribution is based on using a stochastic gradient ascent procedure taking as input a stream of importance sampling weights, so that a mixture of Gaussians is dynamically updated with no need to store the full sample. The algorithm has been designed following a Map/Reduce approach and is therefore scalable with respect to computing resources. The implementation of the proposed algorithm is available as part of the AMIDST open-source toolbox for scalable probabilistic machine learning. (C) 2018 Elsevier Inc. All rights reserved.

引用

页码：115 / 134

页数：20

共 50 条

[1] Improved marginal likelihood estimation via power posteriors and importance sampling
Li, Yong
Wang, Nianling
Yu, Jun
[J]. JOURNAL OF ECONOMETRICS, 2023, 234 (01) : 28 - 52
[2] On the use of marginal posteriors in marginal likelihood estimation via importance sampling
Perrakis, Konstantinos
Ntzoufras, Ioannis
Tsionas, Efthymios G.
[J]. COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2014, 77 : 54 - 69
[3] Probabilistic inference using linear Gaussian importance sampling for hybrid Bayesian networks
Sun, W
Chang, KC
[J]. Signal Processing, Sensor Fusion, and Target Recognition XIV, 2005, 5809 : 322 - 329
[4] Free energy methods for Bayesian inference: efficient exploration of univariate Gaussian mixture posteriors
Chopin, Nicolas
Lelievre, Tony
Stoltz, Gabriel
[J]. STATISTICS AND COMPUTING, 2012, 22 (04) : 897 - 916
[5] Free energy methods for Bayesian inference: efficient exploration of univariate Gaussian mixture posteriors
Nicolas Chopin
Tony Lelièvre
Gabriel Stoltz
[J]. Statistics and Computing, 2012, 22 : 897 - 916
[6] Bayesian estimation of the Gaussian mixture GARCH model
Concepcion Ausin, Maria
Galeano, Pedro
[J]. COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2007, 51 (05) : 2636 - 2652
[7] Bayesian Optimized Mixture Importance Sampling for High-Sigma Failure Rate Estimation
Weller, Dennis D.
Hefenbrock, Michael
Golanbari, Mohammad S.
Beigl, Michael
Aghassi-Hagmann, Jasmin
Tahoori, Mehdi B.
[J]. IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS, 2020, 39 (10) : 2772 - 2783
[8] Sampling of Bayesian posteriors with a non-Gaussian probabilistic learning on manifolds from a small dataset
Soize, Christian
Ghanem, Roger G.
Desceliers, Christophe
[J]. STATISTICS AND COMPUTING, 2020, 30 (05) : 1433 - 1457
[9] MIXED BAYESIAN NETWORKS - A MIXTURE OF GAUSSIAN DISTRIBUTIONS
CHEVROLAT, JP
RUTIGLIANO, F
GOLMARD, JL
[J]. METHODS OF INFORMATION IN MEDICINE, 1994, 33 (05) : 535 - 542
[10] Sampling of Bayesian posteriors with a non-Gaussian probabilistic learning on manifolds from a small dataset
Christian Soize
Roger G. Ghanem
Christophe Desceliers
[J]. Statistics and Computing, 2020, 30 : 1433 - 1457

← 1 2 3 4 5 →