Homotopy continuation for the spectra of persistent Laplacians (Q2072664)

The Laplacian is a critical concept in network analysis and shape tasks. Given its limitation to a single scale of the data set, it was recently imbued with ideas from persistent homology, leading to the \(p\)-persistent \(q\)-combinatorial Laplacian of a simplicial complex. This construction, colloquially referred to as a \textit{persistent Laplacian} can be shown to be intricately linked to multi-scale topological information, such as the persistent Betti numbers of a data set. Motivated by the utility of persistent Laplacians, this paper presents a new way of calculating their spectral information, specifically, the roots of their characteristic polynomials, by homotopy continuation. Homotopy continuation addresses the problem of solving a system~\(f\) of polynomial equations by starting with an easy-to-solve system~\(g\), building a homotopy between~\(f\) and~\(g\), and, finally, tracking the roots of~\(g\) to those of~\(f\). The authors demonstrate the feasibility of this continuation method and provide empirical evidence of the utility of using spectral information from persistent Laplacians to characterise aromatic molecules, for instance.