清华主页 EN
导航菜单

Fast algorithms for hierarchically structured matrices

来源: 08-24

时间:10:15–11:15am, August 26th (Fri), 2022

地点:Zoom Meeting ID: 271 534 5558 Passcode: YMSC

主讲人:Manas Rachh

Abstract

In this talk, I will discuss what hierarchically structured matrices are, where they occur in practice, and present algorithms for factorizing these structured matrices. I will demonstrate how the factorization enables subsequent matrix operations (applying the matrix, computing its inverse, and its determinant) in O(N) times for N x N matrices.


Speaker

Manas Rachh joined the Simons foundation as part of the Numerical Algorithms group at Flatiron’s Center for Computational Biology in 2018, and is currently a research scientist in the Center for Computational Mathematics. His research interests include partial differential equations (PDEs) arising in mathematical physics, integral equation methods, robust computation of eigenvalues and eigenfunctions of elliptic PDEs, and the development of fast algorithms for applications in electrostatics, acoustics, viscous flow, electromagnetics, biomedical imaging, and data visualization. Before coming to the foundation, he obtained his Ph.D. from the Courant Institute of Mathematical Sciences at New York University with Leslie Greengard followed by a Gibbs Assistant Professorship at Yale University where he worked with Vladimir Rokhlin.


返回顶部
相关文章
  • Algorithms in Natural Language Processing

    Record: YesLevel: GraduateLanguage: ChinesePrerequisiteComputer Science, Machine Learning, PythonAbstractNatural Language Processing (NLP) is an important research area in Artificial Intelligence. NLP mainly studys how to use computer technology to process linguistic texts. The specific research problems in NLP includes recognition, classification, extraction, transformation and generation of l...

  • A brief introduction to convergence acceleration algorithms

    AbstractIn this talk, I will give a brief introduction to convergence acceleration algorithms. First, I will provide a concise overview of the research context and the current state of convergence acceleration algorithms. Subsequently, I will present a method known as the kernel methodology for constructing convergence acceleration algorithms, based on which I propose a novel acceleration algor...