Registered Data
Contents
- 1 [CT009]
- 1.1 [00922] Matrix Factorization for Change Detection in HyperSpectral Images
- 1.2 [00546] Recent advances on the conjugate discrete-time algebraic Riccati equation.
- 1.3 [01097] Versal deformations as a tool of matrix analysis
- 1.4 [00206] Tensor product-type methods for solving Sylvester tensor equations
- 1.5 [02254] Optimized first order alternating algorithms for fast and accurate low rank tensor decomposition
[CT009]
[00922] Matrix Factorization for Change Detection in HyperSpectral Images
- Session Date & Time : 3E (Aug.23, 17:40-19:20)
- Type : Contributed Talk
- Abstract : When hyperspectral images are analyzed, a big amount of data needs to be processed and therefore, specific matrix factorization algorithms are used to express the original problem in suitable subspacesWe show some recent results derived also by using spatial and spectral functions to compute a lower rank approximation of the original matrix and to measure the reconstruction error between the input image and the approximate one, with applications to the task of change-detection.
- Classification : 15A23
- Author(s) :
- Antonella Falini (Università degli studi di Bari Aldo Moro)
- Francesca Mazzia (Università degli studi di Bari Aldo Moro, Italy)
[00546] Recent advances on the conjugate discrete-time algebraic Riccati equation.
- Session Date & Time : 3E (Aug.23, 17:40-19:20)
- Type : Contributed Talk
- Abstract : In this talk, we consider a class of conjugate discrete-time Riccati equations, arising originally from the linear quadratic regulation problem for discrete-time antilinear systems. A constructive proof is given for the existence of the maximal solution. Furthermore, an accelerated fixed-point iteration based on the semigroup property is developed for computing the maximal solution and the convergences is at least R-superlinear. An example is given to demonstrate the correctness of our main theorem.
- Classification : 15A24, 65H05, 93A99
- Author(s) :
- Chun-Yueh Chiang (Center for General Education, National Formosa University)
- Hung-Yuan Fan (National Taiwan Normal University)
[01097] Versal deformations as a tool of matrix analysis
- Session Date & Time : 3E (Aug.23, 17:40-19:20)
- Type : Contributed Talk
- Abstract : Reductions of matrices or matrix pencils to canonical forms are unstable operations: both the corresponding canonical forms and the reduction transformations depend discontinuously on the entries of an original matrix or pencil. This issue complicates the use of canonical forms for numerical purposes. Therefore V.I. Arnold introduced a notion of versal deformations. We will discuss versal deformations and their use in codimension computations, investigation of possible changes in eigenstructures, and reduction to structured perturbations.
- Classification : 15A63, 15A21
- Author(s) :
- Andrii Dmytryshyn (Örebro University)
- Andrii Dmytryshyn (Örebro University)
[00206] Tensor product-type methods for solving Sylvester tensor equations
- Session Date & Time : 3E (Aug.23, 17:40-19:20)
- Type : Contributed Talk
- Abstract : The tensor biconjugate gradient $($TBiCG$)$ method has recently been proposed for solving Sylvester tensor equations. TBiCG is based on BiCG that may exhibit irregular convergence behavior. To overcome the limitations of BiCG, product-type methods have been proposed. In this study, we propose tensor product-type methods to solve Sylvester tensor equations. Furthermore, we consider the preconditioned versions using the NKP preconditioner. Numerical experiments illustrate that the proposed methods are competitive with some existing methods.
- Classification : 15A69, 65F10
- Author(s) :
- Jing Niu (Nagoya University)
- Tomohiro Sogabe (Nagoya University)
- Lei Du (Dalian University of Technology)
- Tomoya Kemmochi (Nagoya University)
- Shao-Liang Zhang (Nagoya University)
[02254] Optimized first order alternating algorithms for fast and accurate low rank tensor decomposition
- Session Date & Time : 3E (Aug.23, 17:40-19:20)
- Type : Contributed Talk
- Abstract : CP tensor decomposition has been proven to be a powerful tool for extracting information from large high order tensor, being widely applied in many areas such as chemistry, biology and medical science. However, efficiently computing the CP tensor still remains a challenge. In this study, we propose some optimized first order alternating least square algorithms for low rank tensor decomposition. We validate and illustrate the proposed algorithms by using simulated and real multi-way data.
- Classification : 15A72, 15A69, 65Z05
- Author(s) :
- HUIWEN YU (Aarhus University)
- Ove Christiansen (Aarhus University)