Approximate inference in hidden Markov models using iterative active state selection

CM Vithanage; C Andrieu; RJ Piechocki

doi:10.1109/LSP.2005.861600

Approximate inference in hidden Markov models using iterative active state selection

CM Vithanage, C Andrieu, RJ Piechocki

Research output: Contribution to journal › Article (Academic Journal) › peer-review

5 Citations (Scopus)

464 Downloads (Pure)

Abstract

The inferential task of computing the marginal posterior probability mass functions of state variables and pairs of consecutive state variables of a hidden Markov model is considered. This can be exactly and efficiently performed using a message passing scheme such as the Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm. We present a novel iterative reduced complexity variation of the BCJR algorithm that uses reduced support approximations for the forward and backward messages, as in the M-BCJR algorithm. Forward/backward message computation is based on the concept of expectation propagation, which results in an algorithm similar to the M-BCJR algorithm with the active state selection criterion being changed from the filtered distribution of state variables to beliefs of state variables. By allowing possibly different supports for the forward and backward messages, we derive identical forward and backward recursions that can be iterated. Simulation results of application for trellis-based equalization of a wireless communication system confirm the improved performance over the M-BCJR algorithm

Translated title of the contribution	Approximate inference in hidden Markov models using iterative active state selection
Original language	English
Pages (from-to)	65 - 68
Number of pages	4
Journal	IEEE Signal Processing Letters
Volume	13
Issue number	2
DOIs	https://doi.org/10.1109/LSP.2005.861600
Publication status	Published - Feb 2006

Bibliographical note

Publisher: Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Rose publication type: Journal article

Terms of use: Copyright © 2005 IEEE. Reprinted from IEEE Signal Processing Letters. This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of the University of Bristol's products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to [email protected].
By choosing to view this document, you agree to all provisions of the copyright laws protecting it

Keywords

deterministic algorithms
equalizers
hidden Markov models (HMMs)
message passing
state space methods

Access to Document

10.1109/LSP.2005.861600

Vithanage IEEE SigProcLetsAccepted author manuscript, 199 KB

http://hdl.handle.net/1983/731

Handle.net

Persistent link

Cite this

@article{c4b75a998c124e798a9ff3eb8f00d4e0,

title = "Approximate inference in hidden Markov models using iterative active state selection",

abstract = "The inferential task of computing the marginal posterior probability mass functions of state variables and pairs of consecutive state variables of a hidden Markov model is considered. This can be exactly and efficiently performed using a message passing scheme such as the Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm. We present a novel iterative reduced complexity variation of the BCJR algorithm that uses reduced support approximations for the forward and backward messages, as in the M-BCJR algorithm. Forward/backward message computation is based on the concept of expectation propagation, which results in an algorithm similar to the M-BCJR algorithm with the active state selection criterion being changed from the filtered distribution of state variables to beliefs of state variables. By allowing possibly different supports for the forward and backward messages, we derive identical forward and backward recursions that can be iterated. Simulation results of application for trellis-based equalization of a wireless communication system confirm the improved performance over the M-BCJR algorithm",

keywords = "deterministic algorithms, equalizers, hidden Markov models (HMMs), message passing, state space methods",

author = "CM Vithanage and C Andrieu and RJ Piechocki",

note = "Publisher: Institute of Electrical and Electronics Engineers, Inc. (IEEE) Rose publication type: Journal article Terms of use: Copyright {\textcopyright} 2005 IEEE. Reprinted from IEEE Signal Processing Letters. This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of the University of Bristol's products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to [email protected]. By choosing to view this document, you agree to all provisions of the copyright laws protecting it ",

year = "2006",

month = feb,

doi = "10.1109/LSP.2005.861600",

language = "English",

volume = "13",

pages = "65 -- 68",

journal = "IEEE Signal Processing Letters",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

number = "2",

}

TY - JOUR

T1 - Approximate inference in hidden Markov models using iterative active state selection

AU - Vithanage, CM

AU - Andrieu, C

AU - Piechocki, RJ

N1 - Publisher: Institute of Electrical and Electronics Engineers, Inc. (IEEE) Rose publication type: Journal article Terms of use: Copyright © 2005 IEEE. Reprinted from IEEE Signal Processing Letters. This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of the University of Bristol's products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to [email protected]. By choosing to view this document, you agree to all provisions of the copyright laws protecting it

PY - 2006/2

Y1 - 2006/2

N2 - The inferential task of computing the marginal posterior probability mass functions of state variables and pairs of consecutive state variables of a hidden Markov model is considered. This can be exactly and efficiently performed using a message passing scheme such as the Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm. We present a novel iterative reduced complexity variation of the BCJR algorithm that uses reduced support approximations for the forward and backward messages, as in the M-BCJR algorithm. Forward/backward message computation is based on the concept of expectation propagation, which results in an algorithm similar to the M-BCJR algorithm with the active state selection criterion being changed from the filtered distribution of state variables to beliefs of state variables. By allowing possibly different supports for the forward and backward messages, we derive identical forward and backward recursions that can be iterated. Simulation results of application for trellis-based equalization of a wireless communication system confirm the improved performance over the M-BCJR algorithm

AB - The inferential task of computing the marginal posterior probability mass functions of state variables and pairs of consecutive state variables of a hidden Markov model is considered. This can be exactly and efficiently performed using a message passing scheme such as the Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm. We present a novel iterative reduced complexity variation of the BCJR algorithm that uses reduced support approximations for the forward and backward messages, as in the M-BCJR algorithm. Forward/backward message computation is based on the concept of expectation propagation, which results in an algorithm similar to the M-BCJR algorithm with the active state selection criterion being changed from the filtered distribution of state variables to beliefs of state variables. By allowing possibly different supports for the forward and backward messages, we derive identical forward and backward recursions that can be iterated. Simulation results of application for trellis-based equalization of a wireless communication system confirm the improved performance over the M-BCJR algorithm

KW - deterministic algorithms

KW - equalizers

KW - hidden Markov models (HMMs)

KW - message passing

KW - state space methods

U2 - 10.1109/LSP.2005.861600

DO - 10.1109/LSP.2005.861600

M3 - Article (Academic Journal)

VL - 13

SP - 65

EP - 68

JO - IEEE Signal Processing Letters

JF - IEEE Signal Processing Letters

IS - 2

ER -

Approximate inference in hidden Markov models using iterative active state selection

Abstract

Bibliographical note

Keywords

Access to Document

Handle.net

Fingerprint

Cite this