Online variational message passing in hierarchical autoregressive models

Albert Podusenko; Wouter M. Kouw; A. (Bert) de Vries

doi:10.1109/ISIT44484.2020.9174134

Online variational message passing in hierarchical autoregressive models

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

Hierarchical autoregressive (AR) models can describe many complex physical processes. Unfortunately, online adaptation in these models under non-stationary conditions remains a challenge. In this paper, we track states and parameters in a hierarchical AR filter by means of variational message passing (VMP) in a factor graph. We derive VMP update rules for an "AR node" that can be re-used at various hierarchical levels and supports automated message passing-based inference for states and parameters. The proposed method is experimentally validated for a 2-level hierarchical AR model.

Original language	English
Title of host publication	2020 IEEE International Symposium on Information Theory, ISIT 2020 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers
Pages	1337-1342
Number of pages	6
ISBN (Electronic)	978-1-7281-6432-8
DOIs	https://doi.org/10.1109/ISIT44484.2020.9174134
Publication status	Published - 24 Aug 2020
Event	2020 IEEE International Symposium on Information Theory, ISIT 2020 - Los Angeles, United States Duration: 21 Jun 2020 → 26 Jun 2020

Conference

Conference	2020 IEEE International Symposium on Information Theory, ISIT 2020
Country/Territory	United States
City	Los Angeles
Period	21/06/20 → 26/06/20

Keywords

Variational Message Passing
Autoregressive models
Factor graphs

Access to Document

10.1109/ISIT44484.2020.9174134

https://biaslab.github.io/pdf/isit2020/a_podusenko_online_har.pdf

Cite this

@inproceedings{f49c3eba87dd4d969795435be9701046,

title = "Online variational message passing in hierarchical autoregressive models",

abstract = "Hierarchical autoregressive (AR) models can describe many complex physical processes. Unfortunately, online adaptation in these models under non-stationary conditions remains a challenge. In this paper, we track states and parameters in a hierarchical AR filter by means of variational message passing (VMP) in a factor graph. We derive VMP update rules for an {"}AR node{"} that can be re-used at various hierarchical levels and supports automated message passing-based inference for states and parameters. The proposed method is experimentally validated for a 2-level hierarchical AR model.",

keywords = "Variational Message Passing, Autoregressive models, Factor graphs",

author = "Albert Podusenko and Kouw, \{Wouter M.\} and \{de Vries\}, \{A. (Bert)\}",

year = "2020",

month = aug,

day = "24",

doi = "10.1109/ISIT44484.2020.9174134",

language = "English",

pages = "1337--1342",

booktitle = "2020 IEEE International Symposium on Information Theory, ISIT 2020 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "2020 IEEE International Symposium on Information Theory, ISIT 2020 ; Conference date: 21-06-2020 Through 26-06-2020",

}

Podusenko, A , Kouw, WM & de Vries, A 2020, Online variational message passing in hierarchical autoregressive models. in 2020 IEEE International Symposium on Information Theory, ISIT 2020 - Proceedings., 9174134, Institute of Electrical and Electronics Engineers, pp. 1337-1342, 2020 IEEE International Symposium on Information Theory, ISIT 2020, Los Angeles, California, United States, 21/06/20. https://doi.org/10.1109/ISIT44484.2020.9174134

Online variational message passing in hierarchical autoregressive models. / Podusenko, Albert ; Kouw, Wouter M.; de Vries, A. (Bert).
2020 IEEE International Symposium on Information Theory, ISIT 2020 - Proceedings. Institute of Electrical and Electronics Engineers, 2020. p. 1337-1342 9174134.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Online variational message passing in hierarchical autoregressive models

AU - Podusenko, Albert

AU - Kouw, Wouter M.

AU - de Vries, A. (Bert)

PY - 2020/8/24

Y1 - 2020/8/24

N2 - Hierarchical autoregressive (AR) models can describe many complex physical processes. Unfortunately, online adaptation in these models under non-stationary conditions remains a challenge. In this paper, we track states and parameters in a hierarchical AR filter by means of variational message passing (VMP) in a factor graph. We derive VMP update rules for an "AR node" that can be re-used at various hierarchical levels and supports automated message passing-based inference for states and parameters. The proposed method is experimentally validated for a 2-level hierarchical AR model.

AB - Hierarchical autoregressive (AR) models can describe many complex physical processes. Unfortunately, online adaptation in these models under non-stationary conditions remains a challenge. In this paper, we track states and parameters in a hierarchical AR filter by means of variational message passing (VMP) in a factor graph. We derive VMP update rules for an "AR node" that can be re-used at various hierarchical levels and supports automated message passing-based inference for states and parameters. The proposed method is experimentally validated for a 2-level hierarchical AR model.

KW - Variational Message Passing

KW - Autoregressive models

KW - Factor graphs

UR - https://www.scopus.com/pages/publications/85090414845

U2 - 10.1109/ISIT44484.2020.9174134

DO - 10.1109/ISIT44484.2020.9174134

M3 - Conference contribution

SP - 1337

EP - 1342

BT - 2020 IEEE International Symposium on Information Theory, ISIT 2020 - Proceedings

PB - Institute of Electrical and Electronics Engineers

T2 - 2020 IEEE International Symposium on Information Theory, ISIT 2020

Y2 - 21 June 2020 through 26 June 2020

ER -

Online variational message passing in hierarchical autoregressive models

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this