Motivated by applications in machine learning we consider the notion of total variation defined on graphs over data clouds. Graphs are used to leverage the geometry of a ground-truth distribution, as well as to incorporate "must link" and "can not link" constraints on the data. We study the variational limit of the graph total variation as the number of data points goes to infinity and the parameters used to construct the graphs are scaled appropriately.

This presentation is part of Minisymposium “MS39 - Nonlinear Spectral Theory and Applications (part 2)”
organized by: Aujol Jean-Francois (University of Bordeaux) , Gilboa Guy (Electrical Engineering Department, Technion) .

Authors:

Garcia Trillos Nicolas (Brown University)

Keywords: