Graph Limits via Quotients

Abstract: We present a new notion of limits of weighted directed graphs of growing size based on convergence of their random quotients. These limits are specified in terms of random exchangeable measures on the unit square. We call our limits grapheurs and show that these are dual to graphons in a precise sense. Grapheurs are well-suited to modeling hubs and connections between them in large graphs; previous notions of graph limits based on subgraph densities fail to adequately model such global structures as subgraphs are inherently local. Using our framework, we present an edge-based sampling approach for testing properties pertaining to hubs in large graphs. This method relies on an edge-based analog of the Szemerédi regularity lemma, whereby we show that sampling a small number of edges from a large graph approximately preserves its quotients. Finally, we observe that the random quotients of a graph are related to each other by equipartitions, and we conclude with a characterization of such random graph models.