Abstract:

Massive social and information networks that emerged from online computing
applications have interesting and many times surprising structures. 
Network communities and clusters represent basic building blocks for 
understanding the global organization of such networks. In this talk I will 
address how to extract clusters in massive networks and build understanding 
about the global organization of large networks. Surprisingly, massive online 
social and information networks tend to have a very different community 
structure from well studied small social networks. As small networks often 
exhibit clear community structure, large networks exhibit core-periphery like
structure, with an intermingled network core and many small communities
around it.
While existing network models fail, a simple mathematical model based on
Kronecker products is able to model the organization of large networks. 
Our results suggest broader implications for data analysis in sparse and 
noisy high-dimensional datasets and provide a connection to Dunbar's number 
as a cognitive limit to the size of human communities.

Bio:
Jure Leskovec is an Assistant Professor of Computer Science at Stanford
University. His research focuses on mining and modeling large social and
information networks, their evolution, and diffusion of information and
influence over them. Problems he investigates are motivated by large scale
data, the Web and on-line media. He received three best paper awards, a 
ACM KDD dissertation award, won the ACM KDD Cup in 2003 and topped the 
Battle of the Sensor Networks competition. Jure also holds three patents.