Bayesian uncertainty quantification for low-rank matrix completion

DOI10.1214/22-BA1317MaRDI QIDQ6122033zbMATH OpenOpenAlexFDO

Authors Henry Shaowu Yuchi, Simon Mak, Yao Xie

Publication date 27 February 2024

Published in Bayesian Analysis (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2101.01299, https://projecteuclid.org/journals/bayesian-analysis/volume-18/issue-2/Bayesian-Uncertainty-Quantification-for-Low-Rank-Matrix-Completion/10.1214/22-BA1317.full

zbMATH Keywords

hierarchical modelingmanifold samplingmatrix completionmatrix factorizationseismic imaginguncertainty quantification

Mathematics Subject Classification ID

Statistics (62-XX)

Abstract: We consider the problem of uncertainty quantification for an unknown low-rank matrix $mathbfX$, given a partial and noisy observation of its entries. This quantification of uncertainty is essential for many real-world problems, including image processing, satellite imaging, and seismology, providing a principled framework for validating scientific conclusions and guiding decision-making. However, existing literature has mainly focused on the completion (i.e., point estimation) of the matrix $mathbfX$, with little work on investigating its uncertainty. To this end, we propose in this work a new Bayesian modeling framework, called BayeSMG, which parametrizes the unknown $mathbfX$ via its underlying row and column subspaces. This Bayesian subspace parametrization enables efficient posterior inference on matrix subspaces, which represents interpretable phenomena in many applications. This can then be leveraged for improved matrix recovery. We demonstrate the effectiveness of BayeSMG over existing Bayesian matrix recovery methods in numerical experiments, image inpainting, and a seismic sensor network application.

Cites work

• A Bayesian approach for noisy matrix completion: optimal rate under general sampling distribution
• A simpler approach to matrix completion
• A Singular Value Thresholding Algorithm for Matrix Completion
• Communications-inspired projection design with application to compressive sensing
• Depth Image Inpainting: Improving Low Rank Matrix Completion With Low Gradient Regularization
• Design and analysis of computer experiments. With comments and a rejoinder by the authors
• Equation of state calculations by fast computing machines
• Exact matrix completion via convex optimization
• scientific article; zbMATH DE number 3555239 (Why is no real title available?)
• Inference and uncertainty quantification for noisy matrix completion
• Inference from iterative simulation using multiple sequences
• Matrix Completion From a Few Entries
• Matrix Completion With Deterministic Pattern: A Geometric Perspective
• Model Averaging and Dimension Selection for the Singular Value Decomposition
• Monte Carlo Simulation on the Stiefel Manifold via Polar Expansion
• Nonconvex Optimization Meets Low-Rank Matrix Factorization: An Overview
• Nuclear-norm penalization and optimal rates for noisy low-rank matrix completion
• On block coherence of frames
• On the singular values of Gaussian random matrices
• Regularization and Variable Selection Via the Elastic Net
• Restricted strong convexity and weighted matrix completion: optimal bounds with noise
• Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
• Statistical analysis with missing data
• Support points
• The Power of Convex Relaxation: Near-Optimal Matrix Completion
• Variational Bayesian Super Resolution

Cited in

(2)

• Uncertainty quantification in Bayesian reduced-rank sparse regressions
• Additive Multi-Index Gaussian Process Modeling, with Application to Multi-Physics Surrogate Modeling of the Quark-Gluon Plasma

This page was built for publication: Bayesian uncertainty quantification for low-rank matrix completion

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6122033)