Matthew Hasenjager, University of Tennessee

Abstract

In many complex systems, interactions often involve more than two participants. Such higher-order interactions can have important consequences for understanding processes within these systems. For example, in communication networks, signals can be broadcast to multiple recipients at once or be intercepted by third-party entities (e.g., eavesdropping), highlighting the need to consider higher-order components of network structure to promote desired outcomes or counteract system vulnerabilities. When seeking to model or analyze systems containing higher-order interactions, graph-based networks can be ill-suited, as they cannot effectively distinguish dyadic from higher-order interactions. Hypernetworks are generalizations of networks that can explicitly encode interactions of any order by defining them as hyperedges that contain one or more nodes, and so provide a means to meet this challenge. I will introduce s-centralities as a means of measuring node centrality in higher-order systems and compare their utility to dyadic centrality metrics. Using real-world data on human interaction patterns across multiple contexts, I will first discuss how s-centralities provide unique information about individuals’ social positions, relative to dyadic metrics. Next, I will use simulations of social transmission processes to compare s-centrality to dyadic metrics in terms of capturing individuals’ influence over the speed of transmission under both simple and complex contagions. I will end by discussing future extensions of this work (e.g., measuring centrality in temporal hypernetworks).

Speaker Bio: Matthew is an Intelligence Community Postdoctoral Fellow at the University of Tennessee, Knoxville under the mentorship of Professor Nina Fefferman in the Department of Ecology and Evolutionary Biology and the National Institute for Modeling Biological Systems. He earned his PhD in Biology from the University of Louisville and went on to study honeybee communication as a postdoctoral researcher at Royal Holloway, University of London. He employs theoretical and empirical approaches to investigate the causes of variation in social network structure and its consequences for collective processes, including social learning, communication, and resource distribution. He has worked on a range of systems, including shoaling fish, honeybees, and the development of ant-inspired algorithms for supply chain management. His fellowship is focused on developing tools for hypernetwork analysis and evaluating their relevance for understanding social contagion in higher-order systems.