Seed Seminar

Mathematics

Limiting Degree Distribution for a Sublinear Preferential Attachment Model with Communities

by

Camille Cazaux

on  November-26-2025, 11:30 ! Livein  IHESfor  60min
Abstract:

For many real-world networks, such as the World Wide Web, the degree distribution follows a power law. It is therefore useful to have simple random graph models whose limiting degree distribution exhibits this same feature. With this motivation, physicists Albert-László Barabási and Réka Albert introduced the preferential attachment model that now bears their name. A further advantage of this model is that it incorporates temporal dynamics: starting from an initial graph \(\mathcal{G}_0\) the graph at time \(n+1\) is obtained from the graph at time \(n\), denoted \(\mathcal{G}_{n}\), by adding a new vertex \(v_{n+1}\). This vertex then attaches to one or several vertices of \(\mathcal{G}_n\) according to a preferential attachment rule, meaning that the probability of connecting to a given vertex of \(\mathcal{G}_n\) is proportional to its degree.

We present an extension of this model in which each vertex of the graph is assigned a community (or type), and in which the preferential attachment is sublinear; that is, the probability of attaching to a vertex \(u\) is proportional to \({\rm deg}(u)^\gamma\), where \(\gamma\) is a parameter taking values in \((0,1)\).

 All talks  Fall '25: random forests and fermionic field theories