Alternating Linear Scheme in a Bayesian Framework for Low-Rank Tensor Approximation

DOI10.1137/20M1386414zbMATH Open1489.15038arXiv2012.11228OpenAlexW3116724564WikidataQ114074119 ScholiaQ114074119MaRDI QIDQ5075697FDOQ5075697

Authors: Clara Menzen, Manon Kok, Kim Batselier 

Publication date: 11 May 2022

Published in: SIAM Journal on Scientific Computing (Search for Journal in Brave)

Abstract: Multiway data often naturally occurs in a tensorial format which can be approximately represented by a low-rank tensor decomposition. This is useful because complexity can be significantly reduced and the treatment of large-scale data sets can be facilitated. In this paper, we find a low-rank representation for a given tensor by solving a Bayesian inference problem. This is achieved by dividing the overall inference problem into sub-problems where we sequentially infer the posterior distribution of one tensor decomposition component at a time. This leads to a probabilistic interpretation of the well-known iterative algorithm alternating linear scheme (ALS). In this way, the consideration of measurement noise is enabled, as well as the incorporation of application-specific prior knowledge and the uncertainty quantification of the low-rank tensor estimate. To compute the low-rank tensor estimate from the posterior distributions of the tensor decomposition components, we present an algorithm that performs the unscented transform in tensor train format.

Full work available at URL: https://arxiv.org/abs/2012.11228

Recommendations

• Efficient alternating least squares algorithms for low multilinear rank approximation of tensors
• Rank Regularization and Bayesian Inference for Tensor Completion and Extrapolation
• Nonconvex Robust Low-Rank Tensor Reconstruction via an Empirical Bayes Method
• Orthogonal low rank tensor approximation: alternating least squares method and its global convergence
• Alternating direction method of multipliers for generalized low-rank tensor recovery
• Optimal low-rank approximations of Bayesian linear inverse problems
• Bayesian tensor regression
• Low-rank nonnegative tensor approximation via alternating projections and sketching
• Bayesian factorizations of big sparse tensors
• Low-rank approximation of tensors

zbMATH Keywords

Bayesian inferencelow-rank approximationtensor decompositiontensor trainalternating linear scheme

Mathematics Subject Classification ID

Large-scale problems in mathematical programming (90C06) Factorization of matrices (15A23) Multilinear algebra, tensor calculus (15A69) Bayesian problems; characterization of Bayes procedures (62C10) Least squares and related methods for stochastic control systems (93E24)

Cites Work

• A constructive arbitrary‐degree Kronecker product decomposition of tensors
• Title not available (Why is that?)
• Analysis of individual differences in multidimensional scaling via an \(n\)-way generalization of ``Eckart-Young decomposition
• Tensor Decompositions and Applications
• Tensor-Train Decomposition
• Equivariant and scale-free Tucker decomposition models
• Alternating Minimal Energy Methods for Linear Systems in Higher Dimensions
• The density-matrix renormalization group in the age of matrix product states
• On local convergence of alternating schemes for optimization of convex problems in the tensor train format
• The alternating linear scheme for tensor optimization in the tensor train format
• Approximation of \(2^d\times2^d\) matrices using tensor decomposition
• Solution of Linear Systems and Matrix Inversion in the TT-Format
• A kernel-based framework to tensorial data analysis
• Bayesian filtering and smoothing
• Tensor Decomposition for Signal Processing and Machine Learning
• Probabilistic Tensor Canonical Polyadic Decomposition With Orthogonal Factors
• Tensor network alternating linear scheme for MIMO Volterra system identification
• Stable, Robust, and Super Fast Reconstruction of Tensors Using Multi-Way Projections
• Nonlinear system modeling and identification using Volterra‐PARAFAC models
• Tensor Networks for Dimensionality Reduction and Large-scale Optimization: Part 1 Low-Rank Tensor Decompositions
• Variants of Alternating Least Squares Tensor Completion in the Tensor Train Format
• Tensor Networks for Dimensionality Reduction and Large-scale Optimization: Part 2 Applications and Future Perspectives

Cited In (1)

• Bayesian Tensor Decomposition for Signal Processing and Machine Learning

This page was built for publication: Alternating Linear Scheme in a Bayesian Framework for Low-Rank Tensor Approximation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5075697)