Skip to main content

Computational Geometry and Topological Data Analysis

Mathematics

Image of Herbert Edelsbrunner

Herbert Edelsbrunner

Professor


1 Career

  • Since 2009 Professor at the Institute of Science and Technology Austria (ISTA)
  • 1999 – 2012 Professor of Mathematics at Duke University, USA
  • 1999 – 2012 Arts and Sciences Professor of Computer Science at Duke University, USA
  • 1996 – 2013 Founder, Director, and Principal at Raindrop Geomagic, Inc., USA
  • 1990 – 1999 Professor of Computer Science, University of Illinois at Urbana-Champaign, USA
  • 1987 – 1990 Associated Professor, University of Illinois at Urbana-Champaign, USA
  • 1985 – 1987 Assistant Professor, University of Illinois at Urbana-Champaign, USA
  • 1981 – 1985 Assistant of Information Processing at the Technical University of Graz, Austria

1.1 Selected Distinctions

  • 2020 Ehrenzeichen für Wissenschaft, Forschung und Kunst des Landes Steiermark
  • 2018 Wittgenstein Prize of the Austrian Science Foundation
  • 2014 Full member of the Austrian Academy of Science
  • 2009 Member of the Academia Europaea
  • 2008 Member of the German Academy of Science (Leopoldina)
  • 2006 Honorary Doctoral degree of the Technical University of Graz
  • 2005 Member of the American Academy of Arts and Science
  • 1999 Sir Eward Youde Memorial Fund, Hong Kong University of Science and Technology
  • 1991 Alan T. Waterman Award of the US National Science Foundation

1.2 Selected Grants

  • 2018 – 2023 ERC 788 183, “Alpha shape theory extended”
  • 2008 – 2010 DARPA HR0011-05-1-0057, “Microstates to macrodynamics: a new mathematics of biology”
  • 2000 – 2005 NFS CCR-00-86013, “Computational geometry for structural biology and bioinformatics”

1.3 Tech Transfer (Patents, Spin-Offs)

N/A


1.4 Selected Publications

  1. H. Edelsbrunner. A. Garber and M. Saghafian. On spheres with kk points inside. In “Proc. 41st Intl. Sympos. Comput. Geom., 2025”, 43:1–43:12.
  2. S. Cultrera di Montesano, O. Draganov, H. Edelsbrunner and M. Saghafian. Chromatic alpha complexes. Found. Data Sci. 8 (2025), 30–62.
  3. H. Edelsbrunner and A. Nikitenko. Average and expected distortion of Voronoi paths and scapes. Discrete Comput. Geom. 73 (2025), 490–499.
  4. H. Edelsbrunner and G. Osang. The multi-cover persistence of Euclidean balls. Discrete Comput. Geom. 65 (2021), 1296–1313.
  5. A. Akopyan and H. Edelsbrunner. The weighted mean curvature derivative of a space-filling diagram. Comput. Math. Biophysics 8 (2020), 51–67.

2 Group Members

  • Irene Pisapia, PhD Student since 2026
  • Yossi Bokor Bleile, Postdoc since 2024
  • Michal Lipinski, Postdoc since 2023
  • Nicolo Zava, Postdoc since 2021

3 Research

The general field of Topological Data Analysis started around 2000 and went through a phase of rapid growth within the fields of Mathematics but also Computer Science. It aims at developing tools from algebraic topology to address the needs of the sciences in the analysis of data, which come in ever growing size and sophistication. The primary concept that powers this field is Persistent Homology, introduced by my group around 2000, but similar ideas can be traced back in the mathematical literature to a paper by Marston Morse near the middle of last century. My group takes a broad approach to this field, working on several fronts, including the study of Erdős-type discrete geometry question, the design of fast algorithms, the development of discrete approach to dynamical systems, the investigation of multi-parameter persistent homology, the use and extension of discrete Morse theory, and the stochastic analysis of all these structures for random data and, in particular, Poisson point processes.


3.1 Current Scientific Projects/Directions

  • Quadratic forms to study the morphology of neuronal dendrites
  • Gromov–Hausdorff and other metrics in topological data analysis
  • Depth posets in combinatorial dynamics
  • Graph codes for multi-parameter persistence
  • Random Euclidean minimum spanning trees and generalizations to higher order
  • The discrete geometry of Brillouin zones and related combinatorial questions

3.2 Research in Focus

A bridge connecting stochastic geometry with the algebraic topology of topological data analysis.

A classic question in stochastic geometry is the expected length of the Euclidean minimum spanning tree of nn points chosen uniformly at random in the unit square. In 1959, Beardwood, Halton, and Hammersley proved that there is a a constant, cc, such that the expected length is cnc\sqrt{n} in the limit, when nn goes to infinity. Perhaps surprisingly, it is not know what cc is, and the best bounds prior to our work were 0.6008 ≤ cc ≤ 0.7072, with the upper bound due to Gilbert from 1965 and the lower bound due to Avram and Bertsimas from 1992. In a paper that is currently under consideration for publication, we improved the lower bound to 0.6289. Incidentally, this also improves the prior best lower bound for the shortest traveling salesman tour through the nn points.

While the improvement of the lower bound is not dramatic, I highlight this achievement because
of the approach, which builds a bridge between stochastic geometry and topological data analysis.
This is a two-way bridge. Results on the expected number of critical simplices of a Poisson–Delaunay mosaic are used to find a short proof of an improved lower bound, and the bounds on the Euclidean minimum spanning tree were extended to bounds for norms of chromatic persistent diagrams.



theme sidebar-arrow-up
Back to Top