We will examine flows through a nonlinear eigenvalue perspective. First we will show how gradient descent can reveal the eigen-structure of the initial condition, with a somewhat similar intuition as in the linear case. We will then discuss preliminary findings related to eigenfunctions of anisotropic diffusion.

This presentation is part of Minisymposium “MS15 - Nonlinear Diffusion: Models, Extensions and Algorithms (2 parts)”
organized by: Ke Chen (University of Liverpool) , Joachim Weickert (Saarland University) , Xue-Cheng Tai (Hong Kong Baptist University) .

Authors:

Gilboa Guy (Electrical Engineering Department, Technion)

Keywords:

gradient descent; anisotropic diffusion; nonlinear spectral theory, nonlinear optimization, partial differential equation models