Search

Workshop

Ripser: Efficient computation of Vietoris–Rips persistence barcodes

  • Ulrich Bauer (TU Munich, Germany)
E1 05 (Leibniz-Saal)

Abstract

I will discuss the efficient computation of the Vietoris–Rips persistence barcode for a finite metric space. The implementation in the newly developed C++ code “Ripser” focuses on memory and time efficiency, outperforming previous software by a factor of more than 40 in computation time and a factor of more than 15 in memory efficiency on typical benchmark examples. The improved computational efficiency is based on a close connection between persistent homology and discrete Morse theory, together with novel algorithmic design principles, carrying out a matrix algorithm without actually storing the entire matrix in memory.

Saskia Gutzschebauch

Max-Planck-Institut für Mathematik in den Naturwissenschaften Contact via Mail

Michael Joswig

Max Planck Institute for Mathematics in the Sciences and TU Berlin

Yue Ren

Max Planck Institute for Mathematics in the Sciences