Statistical inference on networks with dependent edges
Jonathan Stewart, Florida State University.
11th of December 2025
Abstract:Â
The question of how to perform statistical inference on networks with dependent edges is an important question in the field of statistical network analysis. The networks of our world are formed through complex stochastic processes which induce non-trivial dependence among the connections and relationships within the network, and many scientific questions of interest center around dependence. There is a need to develop sound statistical methods and theory to help facilitate scientific exploration and learning on networks with dependent edges through statistical inference. This talk discusses recent foundational advancements in theory towards this goal. Applications of this theory will showcase how we can provide rigorous justification of three different inferential goals, which include the consistent estimation of population models for networks with dependent edges, low-rank structure learning for networks with dependent edges, and non-parametric density estimation for dependent graph statistics, highlighting the broad impact of this theory on providing rigorous foundations for performing statistical inference on networks with dependent edges.