Research Group

Current

Violet Ross '25

Noisy edge-copying hypergraphs with node attributes

Justin Corke '25

Assortativity bounds on union graphs
No matching items

Alums

Bell Luo '24

Fast sampling from configuration-model hypergraphs

Real-life polyadic networks call for intrinsically polyadic network science toolkits to analyze them; null models for polyadic network enable contextualization of network measures and statistical inference. This thesis project focuses on hypergraphs as a polyadic network representation, and the configuration model’s extension to hypergraphs as the null model, to present hypersample, an efficient Python package to sample random hypergraphs from the configuration model through hyperedge shuffles. Emphasis is placed on performance optimizations, in terms of both the practical runtime and algorithmic efficiency. A brief demonstration of how our work allows network scientists to produce a null model of hypergraphs to analyze and draw inference from follows.

Next: M.S. in Computer Science, Yale University

Jamie Hackney '24

Opinion dynamics and ranked-choice voting

Ideological polarization affects the ability of democratic governments to function. Levels of ideological polarization have been rising globally in recent years. Many theories have been put forward about potential causes, but the connection between the election system used and the degree of ideological polarization has been less studied. We present an agent based model to simulate plurality and instant-runoff ranked choice election systems based on Bounded Confidence and Attraction-Repulsion opinion updates. We analyze the long term behavior of these models and quantify the degree of ideological polarization as the variance among agent opinions. We provide a set of parameters that cause plurality election systems with Attraction-Repulsion opinion updates to end with high variance while ranked choice election systems end with near-zero variance. We conduct sweeps between four pairs of model parameters and discuss the phase transitions in variance they produce.

Next: Data Science Internship, United Nations
No matching items