Title: Efficient Algorithms in Massive Graphs

Abstract:

Graphs are ubiquitous data presentations that are usually used 
to model complex relations in a wide variety of applications,
including biochemistry, neurobiology, ecology, social sciences,
and information system. The recent explosion in the number of
scale of these real-world structured data sets has created a need
to efficiently process and analyze these massive graphs. Despite
that graph algorithms are well studied over the past decades,
most graph algorithms require to store the whole representations
of graphs, i.e., adjacency matrices, in the memory and do an
off-line computation. As a result it is difficult with these
algorithms to handle massive graphs of more than 100 million
nodes. Hence recent advance in studying graph algorithms is to
design local, streaming and distributed algorithms. In this talk
we discuss this line of research with our recent results for
counting arbitrary subgraphs in massive graphs, and the load
balancing problem in arbitrary networks.