Graphons, Geometry, and Dynamics: Forward and Inverse Perspectives
\'Agnes Backhausz, Christian Kuehn, Sjoerd van der Niet

TL;DR
This paper investigates how graphons encode geometric information, explores spectral properties distinguishing geometric structures, and connects these insights to network dynamics, revealing limitations of spectral invariants.
Contribution
It introduces explicit examples of isospectral graphons with different geometries and links geometric features of graphons to dynamical stability in network models.
Findings
Constructed isospectral graphons with different underlying geometries.
Showed that spectral properties alone may not distinguish geometric differences.
Connected geometric features of graphons to stability in a continuum Kuramoto model.
Abstract
In this work, we explore the interplay between graph limit theory, the geometry of underlying probability spaces, spectral theory, and network dynamical systems. We investigate two primary questions concerning forward and inverse perspectives: first, whether a graphon retains information about the geometry of the space on which it is defined, and second, whether spectral properties can distinguish graphons that originate from different geometric spaces. To address these questions, we differentiate between combinatorial equivalence and geometric structure, highlighting how these concepts are captured simultaneously by the class of pure graphons. Furthermore, we construct explicit examples of isospectral graphons -- graphons whose integral operators share the same spectrum -- that differ in their underlying geometry. By utilizing the heat kernels of Neumann- and Dirichlet-isospectral…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
