Date: Wednesday 30th of March 2022, 13:30 (BST).
Location: Online MS Teams meeting
‘The quest for better network null models’
by Dr Szabolcs Horvát, Max Planck Institute for the Physics of Complex Systems, Dresden, Germany.
What do the workings of the brain, the progression of an epidemic, and shifts in political opinions have in common? All of these are critically influenced by some underlying connectivity structure, such as neural connections or contacts between people. Network science is concerned with uncovering the relationship between the structure and function of such networks: which structural features are responsible for a given observed behaviour? One of the key tools to answering such questions is network null models. Formulating unbiased and principled null models is however challenging, which is why it is common to see ad-hoc approximation or application-inappropriate approaches.
In this talk, I will explore the problem of network model construction through examples from constrained random graphs and spatial networks, and will highlight the importance of a mathematically rigorous approach.