Research

The field of network science is very diverse, in term of real world applications. Whether natural or artificial, most complex systems involve or can be modelled by networks, e.g. spreading of disease, social media, and neural information in the brain.

Despite this the current methods of network generation, and thus much of our analytical work, still relies on fairly coarse network indicators that have been shown to result in high variability in networks with different generation algorithms (Ritchie, Berthouze, & Kiss, 2014).

These issues have led to, and are in part due to, a current lack of understanding of the relationship (or lack thereof) between higher order structures (thou at a much finer level of description) and thus most commonly used to characterize networks (found to often providing an insufficient description of the networks).

This problem of generating realistic and diverse networks that still share the needed structural characteristics is non –trivial and as the field of network modelling become more and more vital, with increasing prevalence of Big Data, there is a growing need for new techniques to explore these questions.

The aim of this PhD project is thus to develop new tools and methodologies to systematically generate networks that exhibit realistic higher-order structure whilst keeping control of classical indicators. (below shows the current exploration of the space of netwotks we our doing).