[talks] Michelle Girvan [UMD, IAS] talk on Wednesday, 3/25, 4:15pm

[Michael J. Freedman]

Hi all,

Our colloquia speaker this week, Michelle Girvan, will be talking about 
clustering and networks.  Beyond being a high school and college friend 
of mine, Michelle is a statistical physicist by training (her PhD at 
Cornell was with Steve Strogatz and postdoc at the Sante Fe Institute), 
but has been venturing into problems of interest to computer scientists, 
applied mathematicians, and sociologists.

Michelle is an assistant professor at University of Maryland, but has 
been visiting IAS for the past year.  The colloquia will be help at 
4:15pm in the small lecture room (105).

--mike

----

Unveiling Isolated and Layered Communities in Complex Networks
Michelle Girvan, University of Maryland, IAS

In the past decade, a number of studies have focused on the statistical
properties of networked systems such as social networks and the
World-Wide Web.  Researchers have focused on properties common to many
real-world networks such as small-world properties, power-law degree
distributions and network transitivity.  In this talk, I will focus on
another property found in many networks, that of community structure, in
which network nodes are joined together in tightly knit groups, between
which there are only looser connections.  I will discuss algorithms for
discovering community structure in networks, beginning with methods that
strictly partition the nodes into non-overlapping groups.  These methods
work by detecting the boundaries between communities and do not require
the user to input the number of desired communities a priori.   I will
also discuss methods for finding highly overlapping or "layered"
community structure in networks.  In this case, using a combination of
the aforementioned techniques, simulated annealing, and other tools
borrowed from statistical physics, it is possible to find multiple
possible divisions of the network into different communities.  I'll
demonstrate that the algorithms are highly effective at discovering
community structure in both computer-generated and real-world network
data, and show how they can be used to shed light on the sometimes
dauntingly complex structure of networked systems.