SIAM awarded the 2018 SIAM Student Paper Prize to Marya Bazzi for her paper, “Community Detection in Temporal Multilayer Networks, with an Application to Correlation Networks,” published in Multiscale Modeling and Simulation in 2016. Bazzi co-authored the paper with Mason A. Porter, Stacy Williams, Mark McDonald, Daniel J. Fenn, and Sam D. Howison while she was a student at the University of Oxford. She received the award and presented her winning paper at the SIAM Annual Meeting, held July 9-13, 2018 in Portland, Oregon.
The SIAM Student Paper Prize recognizes outstanding scholarship by students in applied mathematics and computing as evidenced in a paper accepted for publication in a SIAM journal. The Student Paper Prize is awarded annually to the student authors of the most outstanding papers accepted by a SIAM journal within the three years preceding the nomination deadline. The award is based solely on the merit and content of the candidate’s contribution to the paper.
Marya received her DPhil in Applied Mathematics from the University of Oxford in 2016. Her thesis focused on community structure in time-dependent networks and its application to correlation networks. She was a postdoctoral scholar at the University of Oxford in 2016 and Head of Analytics at a London-based FinTech in 2017. She is currently a Research Fellow at The Alan Turing Institute (with an affiliation to the University of Warwick) and an Alan Tayler Visiting Research Fellow at the University of Oxford.
Q: Why are you excited about winning the SIAM Student Paper Prize?
A: I am both honored and very excited that my work on community structure in time-dependent networks has won a SIAM Student Paper Prize. Network science is a relatively recent interdisciplinary research area, and it’s great to see it increasingly gain recognition in the mathematics research community.
Q: Could you tell us a bit about the research that won you the prize?
A: The research for this paper deals with community detection in time-dependent networks. Networks in their simplest form are graphs and are used to represent a set of entities (nodes) between which pairwise interactions (edges) exist. Communities are broadly defined as sets of nodes that are “more densely” connected to each other than to nodes in the rest of a network, and the analysis of communities has been useful in a wide variety of applications. In many applications, nodes and/or edges vary in time, and most community-detection approaches account for time-dependence in an ad hoc way or via some form of aggregation. We investigate a method known as “multilayer modularity maximization” (a combinatorial optimization problem). This method can detect communities in a sequence of time-dependent networks in a way that explicitly accounts for temporal dependence. We make observations on the parameters of the method with a focus on correlation networks; we prove some theoretical properties on its set of optimal solutions to better understand how to interpret a resulting partition; and we suggest ways to improve scalable algorithms for solving the combinatorial optimization problem in practice. Furthermore, our results apply to any choice of null model in the modularity function (i.e., not solely to the standard null model). The ability to specify a null model is a desirable (albeit underexploited) feature of the method because it can be explicitly adapted to different applications.
Marya Bazzi (right) received the SIAM Student Paper Prize from SIAM President Nick Higham (left) at the 2018 SIAM Annual Meeting.
Q: What does your research mean to the public?
A: Community structure has led to practical insight in various applications of networks such as social networks, information networks, infrastructure networks, biological networks, financial networks, ecological networks, etc. Many applications are time-dependent and it is thus key to develop methods that can account for temporal dependence between network representations of a system at different times. Our work will hopefully lead to further insight in a wide variety of time-dependent applications, as well as further methodological advancements in the area of community detection in time-dependent networks.
Q: What does participation in SIAM mean to you?
A: I have enjoyed and learnt a lot from attending SIAM conferences, reading papers published in SIAM journals and publishing in SIAM.