Cory Glover
Talk recording
Networks are defined by local, node-level connections that give rise to globally rich structures. The relationship between local and global scales has been a central focus in network science, leading to characteristic network measures such as density, degree heterogeneity, clustering, and community structure. In this dissertation, I investigate how local rules govern the emergence of large-scale network properties and structures.The first project focuses on physical networks, where nodes and links occupy physical space, often leading to entangled configurations. This physicality introduces unique constraints and unexplored questions in network science. One such question is what causes links to entangle in a given network. To investigate this, I introduce a new metric, the average crossing number, to quantify the entanglement of physical networks. I demonstrate how key network characteristics—derived from both the adjacency matrix and node positions—control the degree of entanglement.In the second and third projects, I introduce a new concept to network science: network design. Network design is the investigation of how networks are reproduced exactly from a set of pre-defined components. These components have inherent rules determining how they connect together. Using these rules, I demonstrate theoretical conditions which guarantee a single network will form a set of rules. For cases where these conditions are not met, I introduce the concept of guided assembly, where components are aggregated sequentially to form a target structure. I also develop numerical tools to predict network reproducibility for systems which do not fall into our theoretical conditions. Using this theory, I demonstrate how real world networks are reproduced and develop a machine learning framework to unveil the impact network structural features have on network self-assembly. Finally, I design complex self-assembling structures using the theoretical and numerical findings of these projects.
About the speaker
