Consistency of Maximum Likelihood for Continuous-Space Network Models I

Abstract

A very popular class of models for networks posits that each node is represented by a point in a continuous latent space, and that the probability of an edge between nodes is a decreasing function of the distance between them in this latent space. We study the embedding problem for these models, of recovering the latent positions from the observed graph. Assuming certain natural symmetry and smoothness properties, we establish the uniform convergence of the log-likelihood of latent positions as the number of nodes grows. A consequence is that the maximum likelihood embedding converges on the true positions in a certain information-theoretic sense. Extensions of these results, to recovering distributions in the latent space, and so distributions over arbitrarily large graphs, will be treated in the sequel.

17 pages

References in corpus (3)

• The method of moments and degree distributions for network models
• Bootstrapping Networks with Latent Space Structure
• Rates of Convergence of Spectral Methods for Graphon Estimation

Cited by in corpus (5)

• Bootstrapping Networks with Latent Space Structure
• On the Theoretical Properties of the Network Jackknife
• Estimating Graph Dimension with Cross-validated Eigenvalues
• Dynamic Silos: Increased Modularity in Intra-organizational Communication Networks during the Covid-19 Pandemic
• Trading off Accuracy for Speedup: Multiplier Bootstraps for Subgraph Counts