Colloquium / Seminars
Topic：Large-Scale Matrix Computations for Data Science
Speaker：Prof. Huang, Tsung-Ming
(Department of Mathematics, National Taiwan Normal University)
Date time：From 07/27
Graph Laplacian Eigenvalue Problems (GLEP) appear in many areas, such as spectral clustering, image segmentation, dimensionality reduction, data representation, and complex networks. Computing some smallest positive eigenvalues and the associated eigenvectors of GLEP is a fundamental problem in applications. In this short course, we will introduce how to use the Lanczosmethod and the shift-invert residual Arnoldimethod with a preconditioning conjugate gradient method to compute the target eigenpairsefficiently.