On the Uniqueness of the Canonical Polyadic Decomposition of third-order tensors --- Part II: Uniqueness of the overall decomposition

Abstract

Canonical Polyadic (also known as Candecomp/Parafac) Decomposition (CPD) of a higher-order tensor is decomposition in a minimal number of rank-1 tensors. In Part I, we gave an overview of existing results concerning uniqueness and presented new, relaxed, conditions that guarantee uniqueness of one factor matrix. In Part II we use these results for establishing overall CPD uniqueness in cases where none of the factor matrices has full column rank. We obtain uniqueness conditions involving Khatri-Rao products of compound matrices and Kruskal-type conditions.

28 pages

References in corpus (1)

• On the Uniqueness of the Canonical Polyadic Decomposition of third-order tensors --- Part I: Basic Results and Uniqueness of One Factor Matrix

Cited by in corpus (31)

• Tensor Decomposition for Signal Processing and Machine Learning
• Canonical polyadic decomposition of third-order tensors: reduction to generalized eigenvalue decomposition
• On the Uniqueness of the Canonical Polyadic Decomposition of third-order tensors --- Part I: Basic Results and Uniqueness of One Factor Matrix
• An algorithm for generic and low-rank specific identifiability of complex tensors
• Generic uniqueness conditions for the canonical polyadic decomposition and INDSCAL
• DOA Estimation for Transmit Beamspace MIMO Radar via Tensor Decomposition with Vandermonde Factor Matrix
• On generic identifiability of symmetric tensors of subgeneric rank
• Decoupling Multivariate Polynomials Using First-Order Information
• Double Coupled Canonical Polyadic Decomposition for Joint Blind Source Separation
• Effective criteria for specific identifiability of tensors and forms
• The condition number of join decompositions
• Estimating multivariate latent-structure models
• Cramer-Rao-Induced Bounds for CANDECOMP/PARAFAC tensor decomposition
• A condition number for the tensor rank decomposition
• Dictionary-based Tensor Canonical Polyadic Decomposition
• Tensor rank and entanglement of pure quantum states
• Real identifiability vs complex identifiability
• A generalization of Kruskal's theorem on tensor decomposition
• Personalized Coupled Tensor Decomposition for Multimodal Data Fusion: Uniqueness and Algorithms
• Identifiability of Complete Dictionary Learning
• Uniqueness of Nonnegative Tensor Approximations
• A Normal Form Algorithm for Tensor Rank Decomposition
• Perturbation Analysis for Matrix Joint Block Diagonalization
• Separability of Hermitian Tensors and PSD Decompositions
• Toward a generalization of Kruskal's theorem on tensor decomposition
• Linear dependent subsets of Segre varieties
• An Algebraic Approach to Non-Orthogonal General Joint Block Diagonalization
• PARAFAC2-based Coupled Matrix and Tensor Factorizations with Constraints
• A uniqueness result on the decompositions of a bi-homogeneous polynomial
• Semialgebraic Geometry of Nonnegative Tensor Rank
• Deep Tensor Encoding