SPEAKER: Alfred Hero, University of Michigan
TITLE: “Shortest paths through many points”
ABSTRACT: Shortest paths through a set of points (nodes) arise in applications ranging from operations research to machine learning. When the node attributes are real valued and path length is increasing in Euclidean distance, a stochastic homogenization of the discrete shortest path problem reduces it to a continuous calculus of variations problem as the number of nodes approaches infinity. Such continuum limits can give insight and lead to continuum approximations to solutions of the shortest paths problem.
* * * * * *
DATE: Thursday, June 6th, 4:10pm
LOCATION: MSB 1147, Colloquium Room
REFRESHMENTS: 3:30pm MSB 4110 (4th floor lounge)