Today we had an interesting seminar by Sara based on two of her papers. Thanks for the nice presentation!

Abstract: Network science is a fundamental framework to study and understand complex systems in a large variety of phenomena in disciplines like physics, economics, biology and sociology. Dynamics on networks are usually described by defining a linear dynamical model and the properties of such systems can be derived, for each case, by the spectrum of a specific matrix. The characterization of a linear system by its spectrum is recognized but its not sufficiently precise when the linear operator is non-normal. A linear system, in an arbitrary dimension, is non-normal when its governing matrix does not commute with its conjugate transpose. Although a non-normal system can initially be close to an asympotically stable equilibrium, it can leave this state even when a moderate external perturbation occurs and this effect is more notable once one includes stochastic forces to the model. In this perspective, non-normality is an interesting property to investigate stochastic effects like pattern formation or the emerging of stochastic quasicycles.