EVENT DETAILS AND ABSTRACT

Title: Finding Communities in Networks
Speaker: Julia Gaudio
Speaker Info: Northwestern University
Brief Description:
Special Note:
Abstract:

etworks are used to represent physical, biological, and social systems. Many networks exhibit community structure, meaning that there are two or more groups of nodes which are densely connected. Identifying these communities gives valuable insights about the latent features of the nodes. Community detection has been used in a wide array of applications including online advertising, recommender systems (e.g., Netflix), webpage sorting, fraud detection, and neurobiology. I will present my work on algorithms for community detection in three contexts, each with an underlying probabilistic generative model.

(1) Censored networks: How can we identify communities when some connectivity information is missing?

(2) Higher-order networks: Beyond pairwise relationships

(3) Multiple correlated networks: How can we effectively combine data from multiple networks?

Joint work with: Souvik Dhara, Nirmit Joshi, Elchanan Mossel, Miklós Rácz, Colin Sandon, and Anirudh Sridhar