Persistent Homology
Overview
Persistent homology studies the shape of data across scales by recording when topological features are born and when they vanish, often in persistence diagrams or barcodes. Methods build filtrations from point-cloud distances, image intensities, graphs, or multiple functions, then compare or vectorize the resulting diagrams with distances, landscapes, images, and learned representations. These tools are used in classification and analysis of point clouds, graphs, images, time series, and scientific systems, where shape can reveal structure that ordinary statistics may miss. Current work seeks faster, more noise-robust computation and better ways to integrate topology into learning, while the choice of filtration and representation can strongly affect results and benefits remain task-dependent.
Latest papers
52 papers6 papers in the last four weeks, up 20% on the four weeks before. 0.0% of all new papers.