Electrical measures of importance for network edges

Lecturer : 
Charalampos Mavroforakis, Boston University
Event type: 
HIIT seminar
Event time: 
2014-06-16 13:15 to 14:00
Aalto University, Computer Science Building, lecture hall T2

Measures that quantify the importance of nodes and edges in networks are
valuable in the analysis of various types of data, including social, biological,
computer networks and more. With rare exceptions, the applicability of such
measures to large datasets is hindered by the lack of fast algorithms for their
In this talk, we mainly focus on the importance of edges and discuss three
measures that are based on the electrical representation of a graph. We derive
the algebraic forms of these measures from the basic circuit laws of physics and
describe the algorithms that can be used to compute them. We compare the electrical with the combinatorial view for these measures and show how the former one can allow for tools that speed up computations.
Charalampos is finishing his 3rd year as a PhD student in Boston University. There, he is a member of the Data Management lab, collaborating with Evimaria Terzi and George Kollios. During this summer, he is visiting the Data Mining group as an intern, under the guidance of prof. Aristides Gionis.

Last updated on 9 Jun 2014 by Antti Ukkonen - Page created on 9 Jun 2014 by Antti Ukkonen