Xiaoyan Lu advised by Prof. Boleslaw Szymanski defended today successfully his thesis titled: "Community Detection and its Applications in Understanding Dynamics of Social Networks"

Xiaoyan Lu,advised by Prof. Boleslaw Szymanski, defended today successfully his thesis titled: "Community Detection and its Applications in Understanding Dynamics of Social Networks," The Ph.d. Committed included Professors Malik Magdon-Ismail and Jianxi Gao from Computer Science and Gyorgy Korniss from Physics. The thesis was motivated by the desire to understand emergent social phenomena, which like radicalization, civil unrest, and opinion migration are often described as "nobody saw them coming" because of their explosive dynamics, and yet they have a profound impact on our society. Since these dynamics are significantly affected by the social network topology, there is a strong desire to study the interplay between the emergent dynamics and the underlying social network structure. In this thesis, we focus on the community structure in social networks and study its role in the prediction of emergent network dynamics.

Community detection aims at discovering the partition of the network nodes into groups such that the edges inside each group are generally more numerous than the edges across them. In this thesis, we uncover the importance of the implicit dependence of modularity, one of the most widely used quality metrics for community structure, on the resolution parameter, which is the first result connecting the resolution limit of modularity with the random graph models. To avoid the resolution limit in modularity-based community detection, we propose a progressive agglomerative heuristic algorithm based on statistical hypothesis testing to partition a graph recursively, and a semi-supervised weighting scheme to convert the local edge features into edge weights.

Community structure plays an essential role in facilitating the local spread of information. Based on the survival analysis, we formulate a virality prediction framework for global online news using the community structure of online news media. In this thesis, we will introduce how this model classifies the viral news at the early stage of its spread and how we parallelize the corresponding stochastic gradient descent algorithm for the IBM Blue Gene/Q supercomputer at RPI.

Finally, we will discuss the patterns of the polarization evolution found by analyzing millions of roll-call votes in the legislative branches of the United States. We proposed a social dynamical model which explains the directions of polarization change in 28 out of the past 30 U.S. Congresses and derived the tipping points of the dynamical system which provides early warning signals when the system is close to the bifurcation.