A review on network robustness from an information theory perspective

Abstract

The understanding of how a networked system behaves and keeps its topological features when facing element failures is essential in several applications ranging from biological to social networks. In this context, one of the most discussed and important topics is the ability to distinguish similarities between networks. A probabilistic approach already showed useful in graph comparisons when representing the network structure as a set of probability distributions, and, together with the Jensen-Shannon divergence, allows to quantify dissimilarities between graphs. The goal of this article is to compare these methodologies for the analysis of network comparisons and robustness.