Persistent homology is a tool from topology that tracks how geometric features like holes, loops, and voids appear and disappear as you gradually build up a shape from data. Given a cloud of random points sampled from some space, you can construct a nested sequence of geometric objects (like Cech or Vietoris-Rips complexes) and record the "birth" and "death" scales of topological features. The persistence ratio measure summarizes this information statistically. A natural question is whether this summary depends heavily on the specific shape you sampled from and the density of the sampling distribution, or whether it settles into something universal when you have many points.
The paper proves that the answer is universality: as the number of sampled points grows, the average persistence ratio measure converges to a limit that does not depend on which space you sampled from or how the points were distributed across it. Crucially, this holds not just for smooth curved surfaces but for a much broader class of spaces, including shapes with corners, edges, and other singularities, such as those arising in algebraic geometry or piecewise-linear settings. The key geometric condition required is a mild quality assumption ensuring the space is not too wildly curved or folded.
The technical strategy involves approximating complicated singular spaces by simpler triangulated ones using a classical construction called Freudenthal-Kuhn triangulations, then carefully controlling how topological features interact near singular regions where different smooth pieces of the space meet. The result also extends beyond persistent homology to a general family of scale-invariant quantities derived from random geometric complexes. Practically, this universality means that topological statistics extracted from real data carry intrinsic geometric meaning independent of background assumptions, strengthening the theoretical foundations of topological data analysis on non-Euclidean domains.