Abstract
In topological data analysis, persistent homology is used to study the "shapeof data". Persistent homology computations are completely characterized by aset of intervals called a bar code. It is often said that the long intervalsrepresent the "topological signal" and the short intervals represent "noise".We give evidence to dispute this thesis, showing that the short intervalsencode geometric information. Specifically, we prove that persistent homologydetects the curvature of disks from which points have been sampled. We describea general computational framework for solving inverse problems using theaverage persistence landscape, a continuous mapping from metric spaces with aprobability measure to a Hilbert space. In the present application, the averagepersistence landscapes of points sampled from disks of constant curvatureresults in a path in this Hilbert space which may be learned using standardtools from statistical and machine learning.