Dessins d'Enfants as a Topological Lens on Temporal Hypergraphs
Mehmet Emin Aktas, Esra Akbas
Abstract
Hypergraphs are the natural language for group interactions, and topological data analysis has given us several ways to look at them: simplicial and cell complexes, persistent homology, and sheaves. All of these treat a hyperedge as an unordered set. In this paper, we introduce a different and, to our knowledge, new viewpoint that comes from Grothendieck's theory of dessins d'enfants and, unlike these approaches, takes into account the order in which vertices appear in each hyperedge and hyperedges appear around each vertex. A dessin is a bipartite graph drawn on a closed orientable surface. Through Walsh's classical correspondence between hypermaps and bipartite maps, every hypergraph equipped with a cyclic order of incidences at each vertex and inside each hyperedge is a dessin, and therefore determines a surface with faces and a genus. We develop this idea for temporal hypergraphs and build the surface by replaying history: incidences are added in the order in which they occurred, and each new incidence is inserted into the cyclic orders built so far, next to the earlier incidence to which it is related. We prove that every insertion either grows the hypergraph without creating a cycle, closes a cycle locally by splitting a face, or closes it globally by adding a handle; the number of independent cycles equals the number of local plus handle events, and the genus equals the number of handle events. These counts give interpretable, size-controlled measures of how a temporal hypergraph's cycles organize on a surface as it grows. We apply our method to five Stack Exchange communities and show that it separates communities that standard statistics cannot tell apart and quantifies how much the real order of interactions matters. We further discuss what the surface viewpoint adds to existing higher-order network analysis and how it opens a path toward dessin-based learning on hypergraphs.
Create a lesson
Related papers
"I Don't Think It's a Scam": A First Look at Family Intervention Barriers in Elder Abuse Discussions on Reddit and Baidu Tieba
Yipeng Wang, Huy Gia Han Vu, Jiaming Liu et al.
A comparative exploration of Steinerberger curvature and resistance curvature in model and real-world complex networks
Madhumita Mondal, Priyank Vasu, Areejit Samal et al.
Finding Patient Zero in Temporal Contact Networks: A Comparative Study
Nathan Cincera Brack, Martin Sterchi, Lorenz Hilfiker et al.
Inferring Locations and Correcting Bias in Infectious Disease Data Using Human Mobility
Owen Tucker, Lily Wang, Ruochen Kong et al.
Longitudinal Multi-Dimensional Analysis of India-Pakistan Conflict Discourse on Reddit (2019-2025)
Yash Khare, Hunjun Shin, Mohit Singhal
High-Coverage Blind Spots in Online Abuse Monitoring: Account Selection, Held-Out Flagging Rates, and Sample Composition
Ziming Liu