University Calendar
Topological Data Analysis: Applications and Directions
Dr. Francis Motta, Duke University, presents his seminar.
About the Seminar
"What is the shape of your data?" What may have been regarded as a peculiar question at one time, today is being frequently asked and answered by researchers who are searching to leverage latent geometric structure in their data to both improve their models and gain new insights. This talk will serve as an introduction to persistent homology (PH), a topological data analysis tool rooted in algebraic topology, and will highlight examples that showcase the applicability of PH to a diverse set of questions and data types. I will then discuss recent efforts to strengthen the bridge between machine learning and PH, which may be viewed as a tool for extracting shape-based features that are useful for classification and prediction. Finally, I will describe some facets of the future of the field of topological data analysis, especially those related to time-varying data, by sharing some recent numerical and theoretical results.