Hybrid Vector Message Passing for Generalized Bilinear Factorization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jiang, Hao, Yuan, Xiaojun, Guo, Qinghua
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911750339690496
author Jiang, Hao
Yuan, Xiaojun
Guo, Qinghua
author_facet Jiang, Hao
Yuan, Xiaojun
Guo, Qinghua
contents In this paper, we propose a new message passing algorithm that utilizes hybrid vector message passing (HVMP) to solve the generalized bilinear factorization (GBF) problem. The proposed GBF-HVMP algorithm integrates expectation propagation (EP) and variational message passing (VMP) via variational free energy minimization, yielding tractable Gaussian messages. Furthermore, GBF-HVMP enables vector/matrix variables rather than scalar ones in message passing, resulting in a loop-free Bayesian network that improves convergence. Numerical results show that GBF-HVMP significantly outperforms state-of-the-art methods in terms of NMSE performance and computational complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2401_03626
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Hybrid Vector Message Passing for Generalized Bilinear Factorization
Jiang, Hao
Yuan, Xiaojun
Guo, Qinghua
Signal Processing
In this paper, we propose a new message passing algorithm that utilizes hybrid vector message passing (HVMP) to solve the generalized bilinear factorization (GBF) problem. The proposed GBF-HVMP algorithm integrates expectation propagation (EP) and variational message passing (VMP) via variational free energy minimization, yielding tractable Gaussian messages. Furthermore, GBF-HVMP enables vector/matrix variables rather than scalar ones in message passing, resulting in a loop-free Bayesian network that improves convergence. Numerical results show that GBF-HVMP significantly outperforms state-of-the-art methods in terms of NMSE performance and computational complexity.
title Hybrid Vector Message Passing for Generalized Bilinear Factorization
topic Signal Processing
url https://arxiv.org/abs/2401.03626