主 题：Computing convolutions, the story so far
主讲人：University of Kent 徐宽助理教授
Assistant Professor, School of Mathematics, Statistics, and Actuarial Science, University of Kent, UK, August 2015 - Present；
Postdoctoral research fellow, Mathematical Institute, Oxford University, UK, September 2012 - August 2015；
Algorithm Scientist, General Electric Global Research, General Electric, Niskayuna, NY, USA, May 2011 to June 2012
PhD, Computational Mathematics, New Jersey Institute of Technology & Rutgers, the State University of New Jersey, New Jersey, USA, August 2005 - August 2010；
BS, Computational Mathematics, Beijing University of Aeronautics and Astronautics, Beijing, China, October 2001 - July 2005
Convolution abounds in mathematics and engineering. In this talk, we'll review classic results in literature on convolution quadrature before moving on to the most recent development of powerful numerical methods for computing convolution integrals. Based on the spectral approximation of convolution operators via classic orthogonal polynomials or Fourier extensions, we'll arrive at fast and spectrally-accurate algorithms which make the calculation of convolution integrals possible in the spirit of "computing with functions" and these new methods are believed to lay the foundation of the spectral methods for convolution integral equations.