Project 17: "Computational geometry and topology"

This project will start in 2019.

Principal investigator: Michael Kerber
Graz University of Technology, Austria.
Mentor for: Hüning, Del Giudice.

Project description

My research focuses on the area of applied algebraic topology. The theory of persistent homology permits the usage of classical homology theory in the context of noisy data. This has lead to an entirely new discipline named topological data analysis with applications to applied fields such as neuroscience, material science, medical science, and many others. See this survey for a more thorough introduction.

Thanks to the wide-spread application domains of the developed theory, there is an ever-increasing demand for computational methods to apply persistent homology to larger (and high-dimensional) data sets. My primary interest is the development of efficient methods for the computation of topological descriptors applicable to data analysis. Here, ``efficient'' means both theoretically efficient in terms of worst-case asymptotic complexity (connecting to theoretical computer science), and practically efficient in terms of runtime and memory consumption (reaching into algorithm engineering and software development).

I am especially interested in the above questions in the context of geometric input data. Besides being an important special case, the geometric perspective often provides intuition and leads to approaches that can later be generalized to more general inputs. Another aspect of my research is to apply efficient methods from computational geometry in topological contexts. More often than not, the application of such methods requires the development of additional theory in geometry first.

Showcases for possible PhD themes can be found here.