Low-CP-rank tensor completion via practical regularization

DOI10.1007/S10915-022-01789-9zbMATH Open1495.65062arXiv2103.16852OpenAlexW3148396924MaRDI QIDQ2113663FDOQ2113663

C. Navasca, Fatoumata Sanogo, Jiahua Jiang

Publication date: 14 March 2022

Published in: Journal of Scientific Computing (Search for Journal in Brave)

Abstract: Dimension reduction techniques are often used when the high-dimensional tensor has relatively low intrinsic rank compared to the ambient dimension of the tensor. The CANDECOMP/PARAFAC (CP) tensor completion is a widely used approach to find a low-rank approximation for a given tensor. In the tensor model, an $el{l}_1$ regularized optimization problem was formulated with an appropriate choice of the regularization parameter. The choice of the regularization parameter is important in the approximation accuracy. However, the emergence of the large amount of data poses onerous computational burden for computing the regularization parameter via classical approaches such as the weighted generalized cross validation (WGCV), the unbiased predictive risk estimator, and the discrepancy principle. In order to improve the efficiency of choosing the regularization parameter and leverage the accuracy of the CP tensor, we propose a new algorithm for tensor completion by embedding the flexible hybrid method into the framework of the CP tensor. The main benefits of this method include incorporating regularization automatically and efficiently, improved reconstruction and algorithmic robustness. Numerical examples from image reconstruction and model order reduction demonstrate the performance of the propose algorithm.

Full work available at URL: https://arxiv.org/abs/2103.16852

Recommendations

• Recovering low CP/Tucker ranked tensors, with applications in tensor completion
• Fundamental conditions for low-CP-rank tensor completion
• Tensor completion and low-\(n\)-rank tensor recovery via convex optimization
• Tensor completion using total variation and low-rank matrix factorization
• Tensor Factorization for Low-Rank Tensor Completion
• Low-Tubal-Rank Tensor Completion Using Alternating Minimization
• Low-rank tensor completion by Riemannian optimization
• Low-rank tensor completion using matrix factorization based on tensor train rank and total variation
• Low-rank tensor completion based on log-det rank approximation and matrix factorization
• Low-rank tensor completion via smooth matrix factorization

zbMATH Keywords

tensorregularizationmodel order reductiontensor completionhybrid projection methods

Mathematics Subject Classification ID

Multilinear algebra, tensor calculus (15A69) Matrix completion problems (15A83) Numerical methods for low-rank matrix approximation; matrix compression (65F55)

Cites Work

• A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems
• Estimation of the mean of a multivariate normal distribution
• Three-way arrays: rank and uniqueness of trilinear decompositions, with application to arithmetic complexity and statistics
• The Power of Convex Relaxation: Near-Optimal Matrix Completion
• Algorithms for Numerical Analysis in High Dimensions
• Title not available (Why is that?)
• Spectral Methods for Time-Dependent Problems
• Spectral Methods in MATLAB
• Computational Methods for Inverse Problems
• Title not available (Why is that?)
• Local convergence of the alternating least squares algorithm for canonical tensor approximation
• Convergence Rates for Greedy Algorithms in Reduced Basis Methods
• A bidiagonalization algorithm for solving large and sparse ill-posed systems of linear equations
• Title not available (Why is that?)
• Sparse Regularization via Convex Analysis
• Discrete Inverse Problems
• Title not available (Why is that?)
• A Bidiagonalization-Regularization Procedure for Large Scale Discretizations of Ill-Posed Problems
• Some convergence results on the regularized alternating least-squares method for tensor decomposition
• A weighted-GCV method for Lanczos-hybrid regularization
• Krylov methods for inverse problems: Surveying classical, and introducing new, algorithmic approaches
• A dimensional reduction approach based on the application of reduced basis methods in the framework of hierarchical model reduction
• Low‐rank approximation of tensors via sparse optimization
• The projected GSURE for automatic parameter tuning in iterative shrinkage methods
• Tensor methods for the Boltzmann-BGK equation
• Flexible Krylov Methods for $\ell_p$ Regularization
• Linear Support Tensor Machine With LSK Channels: Pedestrian Detection in Thermal Infrared Images

Cited In (5)

• Title not available (Why is that?)
• Iterative hard thresholding for low CP-rank tensor models
• TR-STF: a fast and accurate tensor ring decomposition algorithm via defined scaled tri-factorization
• Low-rank tensor completion using matrix factorization based on tensor train rank and total variation
• Low-Tubal-Rank Tensor Completion Using Alternating Minimization

Uses Software

• Matlab

This page was built for publication: Low-CP-rank tensor completion via practical regularization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2113663)