Low‐rank approximation of tensors via sparse optimization

The goal of this paper is to find a low‐rank approximation for a given nth tensor. Specifically, we give a computable strategy on calculating the rank of a given tensor, based on approximating the solution to an NP‐hard problem. In this paper, we formulate a sparse optimization problem via an l₁‐regularization to find a low‐rank approximation of tensors. To solve this sparse optimization problem, we propose a rescaling algorithm of the proximal alternating minimization and study the theoretical convergence of this algorithm. Furthermore, we discuss the probabilistic consistency of the sparsity result and suggest a way to choose the regularization parameter for practical computation. In the simulation experiments, the performance of our algorithm supports that our method provides an efficient estimate on the number of rank‐one tensor components in a given tensor. Moreover, this algorithm is also applied to surveillance videos for low‐rank approximation.     

Modifiable low‐rank approximation to a matrix

A truncated ULV decomposition (TULVD) of an m×n matrix X of rank k is a decomposition of the form X = ULV^T+E, where U and V are left orthogonal matrices, L is a k×k non‐singular lower triangular matrix, and E is an error matrix. Only U,V, L, and ∥E∥_F are stored, but E is not stored. We propose algorithms for updating and downdating the TULVD. To construct these modification algorithms, we also use a refinement algorithm based upon that in (SIAM J. Matrix Anal. Appl. 2005; 27 (1):198–211) that reduces ∥E∥_F, detects rank degeneracy, corrects it, and sharpens the approximation. Copyright © 2009 John Wiley & Sons, Ltd.     

On the approximation of symmetric operators by operators of finite rank

The following Theorem 1 and Theorem 2 are established. The proof utilizes the elementary properties of symmetric operators. It is shown that the symmetry condition is necessary for these theorems to hold. This work was sponsored in part by NSF contract 2426.     

Block‐row Hankel weighted low rank approximation

This paper extends the weighted low rank approximation (WLRA) approach to linearly structured matrices. In the case of Hankel matrices with a special block structure, an equivalent unconstrained optimization problem is derived and an algorithm for solving it is proposed. Copyright © 2005 John Wiley & Sons, Ltd.     

Rational matrix approximation with a priori error bounds for non-symmetric matrix Riccati equations with analytic coefficients

In this paper the exact solution of the non-symmetric matrixRiccati equation with analytic coefficients is approximatedby a rational matrix function with a prefixed accuracy. Thisrational matrix function is locally defined as the exact solutionof a Riccati problem with matrix polynomial coefficients obtainedby truncation of the Taylor expansions of the matrix coefficientsof the original problem.     

Generalized inversion of finite rank Hankel and Toeplitz operators with rational matrix symbols

The goal of the paper is a generalized inversion of finite rank Hankel operators and Hankel or Toeplitz operators with block matrices having finitely many rows. To attain it a left coprime fractional factorization of a strictly proper rational matrix function and the Bezout equation are used. Generalized inverses of these operators and generating functions for the inverses are explicitly constructed in terms of the fractional factorization.     

A non-linear structure preserving matrix method for the low rank approximation of the Sylvester resultant matrix

A non-linear structure preserving matrix method for the computation of a structured low rank approximation of the Sylvester resultant matrix S(f,g) of two inexact polynomials f=f(y) and g=g(y) is considered in this paper. It is shown that considerably improved results are obtained when f(y) and g(y) are processed prior to the computation of , and that these preprocessing operations introduce two parameters. These parameters can either be held constant during the computation of , which leads to a linear structure preserving matrix method, or they can be incremented during the computation of , which leads to a non-linear structure preserving matrix method. It is shown that the non-linear method yields a better structured low rank approximation of S(f,g) and that the assignment of f(y) and g(y) is important because may be a good structured low rank approximation of S(f,g), but may be a poor structured low rank approximation of S(g,f) because its numerical rank is not defined. Examples that illustrate the differences between the linear and non-linear structure preserving matrix methods, and the importance of the assignment of f(y) and g(y), are shown.     

Truncated low‐rank methods for solving general linear matrix equations

This work is concerned with the numerical solution of large‐scale linear matrix equations . The most straightforward approach computes from the solution of an mn × mn linear system, typically limiting the feasible values of m,n to a few hundreds at most. Our new approach exploits the fact that X can often be well approximated by a low‐rank matrix. It combines greedy low‐rank techniques with Galerkin projection and preconditioned gradients. In turn, only linear systems of size m × m and n × n need to be solved. Moreover, these linear systems inherit the sparsity of the coefficient matrices, which allows to address linear matrix equations as large as m = n = O(10⁵). Numerical experiments demonstrate that the proposed methods perform well for generalized Lyapunov equations. Even for the case of standard Lyapunov equations, our methods can be advantageous, as we do not need to assume that C has low rank. Copyright © 2015 John Wiley & Sons, Ltd.     

A preconditioned low‐rank CG method for parameter‐dependent Lyapunov matrix equations

This paper is concerned with the numerical solution of symmetric large‐scale Lyapunov equations with low‐rank right‐hand sides and coefficient matrices depending on a parameter. Specifically, we consider the situation when the parameter dependence is sufficiently smooth, and the aim is to compute solutions for many different parameter samples. On the basis of existing results for Lyapunov equations and parameter‐dependent linear systems, we prove that the tensor containing all solution samples typically allows for an excellent low multilinear rank approximation. Stacking all sampled equations into one huge linear system, this fact can be exploited by combining the preconditioned CG method with low‐rank truncation. Our approach is flexible enough to allow for a variety of preconditioners based, for example, on the sign function iteration or the alternating direction implicit method. Copyright © 2013 John Wiley & Sons, Ltd.     

On some matrix equalities for generalized inverses with applications

Necessary and sufficient conditions are derived for the matrix equality A^-=PN^-Q to hold, where A^- and N^- are generalized inverses of matrices. Some consequences and applications are also given. In particular, necessary and sufficient conditions are derived for the additive decompositions C^-=A^-+B^- and to hold.     

Asymptotic expansions of matrix coefficients: The real rank one case

Let G be a real reductive Lie group of class $\text{H}$, and suppose that the split rank of G is one. We show that the asymptotic expansions of the Eisenstein integrals given in Harish-Chandra (1) give uniform approximation off of a certain naturally defined compact subset of $\text{A}\text{?}$, the unitary dual of A; G = KAN being an Iwasawa decomposition of G.     

Cores for parabolic operators with unbounded coefficients

Let be a family of elliptic differential operators with unbounded coefficients defined in RN+1. In [M. Kunze, L. Lorenzi, A. Lunardi, Nonautonomous Kolmogorov parabolic equations with unbounded coefficients, Trans. Amer. Math. Soc., in press], under suitable assumptions, it has been proved that the operator G:=A−Ds generates a semigroup of positive contractions (Tp(t)) in Lp(RN+1,ν) for every 1?p<+∞, where ν is an infinitesimally invariant measure of (Tp(t)). Here, under some additional conditions on the growth of the coefficients of A, which cover also some growths with an exponential rate at ∞, we provide two different cores for the infinitesimal generator Gp of (Tp(t)) in Lp(RN+1,ν) for p∈[1,+∞), and we also give a partial characterization of D(Gp). Finally, we extend the results so far obtained to the case when the coefficients of the operator A are T-periodic with respect to the variable s for some T>0.     

Projection methods for computing Moore-Penrose inverses of unbounded operators

In this article we give a characterization of the convergence of projection methods which are useful for approximating the Moore-Penrose inverse of a closed densely defined operator between Hilbert spaces. We illustrate the main theorem with an example. Also a procedure for constructing the admissible sequence of projections is discussed.     

Dual‐ and triple‐mode matrix approximation and regression modelling

We propose a dual‐ and triple‐mode least squares for matrix approximation. This technique applied to the singular value decomposition produces the classical solution with a new interpretation. Applied to regression modelling, this approach corresponds to a regularized objective and yields a new solution with properties of a ridge regression. The results for regression are robust and suggest a convenient tool for the analysis and interpretation of the model coefficients. Numerical results are given for a marketing research data set. Copyright © 2003 John Wiley & Sons, Ltd.     

Discrete operators of sampling type and approximation in ‐variation

The paper deals with approximation results with respect to the φ‐variation by means of a family of discrete operators for φ‐absolutely continuous functions. In particular, for the considered family of operators and for the error of approximation, we first obtain some estimates which are important in order to prove the main result of convergence in φ‐variation. The problem of the rate of approximation is also studied. The discrete operators that we consider are deeply connected to some problems of linear prediction from samples in the past, and therefore have important applications in several fields, such as, for example, in speech processing. Moreover such family of operators coincides, in a particular case, with the generalized sampling‐type series on a subset of the space of the φ‐absolutely continuous functions: therefore we are able to obtain a result of convergence in variation also for the generalized sampling‐type series. Some examples are also discussed.     

On the basis property of the root functions of differential operators with matrix coefficients

We obtain asymptotic formulas for eigenvalues and eigenfunctions of the operator generated by a system of ordinary differential equations with summable coefficients and periodic or antiperiodic boundary conditions. Then using these asymptotic formulas, we find necessary and sufficient conditions on the coefficients for which the system of eigenfunctions and associated functions of the operator under consideration forms a Riesz basis.     

Efficient inversion of the Galerkin matrix of general second-order elliptic operators with nonsmooth coefficients

This article deals with the efficient (approximate) inversion of finite element stiffness matrices of general second-order elliptic operators with -coefficients. It will be shown that the inverse stiffness matrix can be approximated by hierarchical matrices ( -matrices). Furthermore, numerical results will demonstrate that it is possible to compute an approximate inverse with almost linear complexity.