Abstract

Comparing networks is important for understanding the organizational principles and tracking the dynamics in a broad spectrum of areas that involve many interacting objects. Examples include relationships between people, the global climate system, the world’s economic system, bindings between bio-molecules in a cell, and much else. We introduce new network analysis, comparison and alignment methods, based on counts of local subnetworks, which are easy to compute and produce meaningful and robust results along a wide array of networks. We validate them on simple models, for which, unlike for real world, we know the answers and hence can check the validity of the methods. Also, we apply them to real-world networks from several domains and uncover new domain-specific knowledge.

NOTE that the seminar will take place in the Munby Room in King’s College (it’s not far from Keynes Hall and there will be signs to direct you there).