Singular values and eigenvalues of tensors: a variational approach
By Lek-Heng Lim. IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, 2005
1,090 citations recorded by the source corpus.
We propose a theory of eigenvalues, eigenvectors, singular values, and singular vectors for tensors based on a constrained variational approach much like the Rayleigh quotient for symmetric matrix eigenvalues. These notions are particularly useful in generalizing certain areas where the spectral theory of matrices has traditionally played an important role. For illustration, we will discuss a multilinear generalization of the Perron-Frobenius theorem
Subfields: Computational Mathematics, Computational Theory and Mathematics, Signal Processing
View Paper: https://www.semanticscholar.org/paper/689552d6e1f83984c3bd4d0af3bc33304931d1ca
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Mathematics
- Liqun Qi
- Tamara G. Kolda
- Chaoqian Li
- Yaotang Li
- Anima Anandkumar
- Misha Elena Kilmer
- Guanglu Zhou
- Evrim Acar
In Computational Theory and Mathematics
- Liqun Qi
- Chaoqian Li
- Yaotang Li
- Guanglu Zhou
- J. Shao
- Kelly J. Pearson
- L. Caccetta
- Misha Elena Kilmer
In Signal Processing
- Anima Anandkumar
- Tamara G. Kolda
- Evrim Acar
- Rong Ge
- Misha Elena Kilmer
- Boris N. Khoromskij
- Daniel J. Hsu
- Ning Hao
Citations in the corpus
- A new Brauer-type eigenvalue localization set for tensors (2016, background): “Independently, in [4], Lim gave such a definition but restricted x to be a real vector and λ to be a real number.”
- Perron-Frobenius theorem for nonnegative tensors (2008, methodology): “When m is even, and C is symmetric, this was introduced by Qi [7]; when m is odd, Lim [4] used (x m−1 sgn x1 1 , . . . , x m−1 n sgn x n ) on the right-hand side instead, and the notion has been generalized in Chang Pearson Zhang [1].”
- Unsupervised Multiway Data Analysis: A Literature Survey (2009, background): “Later, computation of singular values and singular vectors of tensors has been discussed in depth in [37] (or N-mode SVD in [38] and [39]).”
- Efficient MATLAB Computations with Sparse and Factored Tensors (2007, neutral): “For example, Lim [32] proposes that the covariant aspect of the multiplication be made explicit by expressing (4.”
- Tensor inversion and its application to the tensor equations with Einstein product (2018, background): “Additionally, the following example reveals that the eigenvalue problem (3) is very different from those presented in [15,16].”
- On best rank one approximation of tensors (2011, methodology): “We will make use of the following result of [14], where we present the proof for completeness.”
- Vacuum stability of a general scalar potential of a few fields (2016, background): “Can one generalise matrix eigenvalues and eigenvectors to tensors in such a way that they have similar properties? Indeed, tensor eigenvalues and eigenvectors have been defined independently by Qi in [44] (which we follow in our exposition) and Lim [45].”
- On some properties of three different types of triangular blocked tensors (2016, neutral): “Lim ([11]) in 2005.”
- Spectra of Uniform Hypergraphs (2011, methodology): “Recent work [9, 17, 22, 24, 26] has provided some of the framework and tools with which to analyze such higher dimensional arrays, and we employ these developments extensively to define and analyze our hypergraph eigenvalues.”
- Third-Order Tensors as Operators on Matrices: A Theoretical and Computational Framework with Applications in Imaging (2013, background): “Eigenvalue decomposition for tensors means different things to different researchers (see [12] [16] [18] [19], for example).”
- On the spectrum of hypergraphs (2017, background): “ypergraph on n vertices, where each edge contains the same, m, number of vertices can easily be represented by a hypermatrix of order m and dimension n. In 2005 [33], L. Qi and independently L.H. Lim [26] introduced the concept of eigenvalues of a real supersymmetric tensor (hypermatrix). Then spect”
- Best rank one approximation of real symmetric tensors can be chosen symmetric (2011, background): “[d]xii2 +kP (T )k2. (3.6) Hence a minimal solution of the left-hand side of (1.1) gives rise to a maximum or minimum of hT ,⊗j∈[d]xji on S n1−1×...×Snd−1. We now give a short proof of a result by Lim [9]. Consider the maximum problem (1.2). Use Lagrange multipliers for the function hT ,⊗j∈[d]xji − P”
- Z-EIGENVALUE INCLUSION THEOREMS FOR TENSORS (2016, background): “There is abundant literature on the development of H-eigenvalues for nonnegative tensors [2, 4, 8, 9, 12, 14, 20, 22, 23].”
- Infinite and finite dimensional Hilbert tensors (2014, background): “Lim [9] independently introduced the notion of eigenvalue for higher order tensors but restricted x to be a real vector and λ to be a real number.”
- M-Tensors and Some Applications (2014, positive): “Eigenvalues of tensors were introduced in [17, 22] in 2005.”