Personal tools

Nested Dissection and Persistent Homology

— filed under:

Dr. Don Sheehy, University of Connecticut

  • Topology Seminar
When Mon, Feb 03, 2014
from 11:00 AM to 11:50 AM
Where RH 316
Add event to calendar vCal
Nested dissection is a way of solving systems of linear equations by divide and conquer.  For many linear systems that occur naturally in many settings, nested dissection gives a guaranteed improvement over naive Gaussian elimination.  These are the so-called beta-separable systems, where beta is a constant that governs how big the improvement will be.  In this talk, I will give some historical background, going back to Strassen's fast matrix multiplication algorithm.  Then, I will show how one might apply this technique to computing persistent homology.  Moreover, I will show that for a wide class of inputs that come up in persistent homology, the resulting systems are beta-separable, yielding an improvement in the asymptotic running time of the persistence algorithm.  
This is joint work with Primoz Skraba and Michael Kerber.
« March 2018 »
Upcoming Events
Math/CS Club
Wed, Mar 21, 2018
100 prisoners Bryan Clair, SLU
Fri, Mar 23, 2018
Introduction to Cluster Algebras Kyungyong Lee, University of Nebraska, Lincoln
Wed, Apr 18, 2018
Quantitative Topology? by Shmuel Weinberger, U Chicago
Annual Awards Ceremony
Fri, Apr 27, 2018
The 2018 Math & Statistics Department Award Ceremony Featuring Sarah Greenwald, Appalachian State University
Previous events…
Upcoming events…