Research
Research
My research lies at the intersection of Topology, Combinatorics, and Data Science. It loosely fits under the umbrella of Topological Data Analysis (TDA), a burgeoning field over the past 25 years.
Data is increasingly relevant in our world, but also increasingly complex. My research attempts to deal with this complexity in two ways: the models used to represent the data, and the methods used to analyze it.
Mathematicians have been modeling data problems with graphs for a long time. Think of the travelling salesperson problem, or Euler's Bridges of Königsberg. Graphs are fantastic tools for displaying pairwise information. However, it is often beneficial to model data with relationships of higher order. For example, biologists have found examples of three-way symbiotic relationships. I use combinatorial objects called hypergraphs to model data. Hypergraphs are generalizations of graphs in which edges can have any number of vertices. The image above shows a hypergraph with various edge sizes, visualized using the python package HypernetX.
Now that we have slightly more sophisticated models for data in hypergraphs, we need more sophisticated methods to analyze those hypergraphs as well. Homology is a tool from Algebraic Topology that captures information about the shape of topological spaces. Hypergraphs are also generalizations of simplicial complexes, which have a well understood homology theory. Much of my current research focuses on adapting simplicial homology theory to hypergraphs.
From February to August 2022, I was privileged to have an internship with Pacific Northwest National Laboratory (PNNL). While there, I learned about research in a nonacademic setting, current applications of the theory in my field, and collaborative research. While here, I collaborated on a paper for the ICML Workshop on Machine Learning for Cybersecurity. You can also view the paper.
While at PNNL, I also wrote code to compute two of the hypergraph homology theories that I focus on in my research. Having this code available makes data analysis more accessible.
While studying at Bowling Green State University, I was a research assistant for Dr. Kimberly Rogers. We were studying the effects of a peer mentoring system that BGSU used for their new graduate student instructors. This research experience resulted in an article for the Journal of Mathematics Teacher Education.
Christopher Potvin - christopherd.potvin@gmail.com - (443) 859-2483