Title:Slow-Down vs. Speed-Up of Information Diffusion in Non-Markovian Temporal Networks

Abstract:We study the slow-down and speed-up of information diffusion in temporal networks with non-Markovian contact sequences. We introduce a causality-preserving time-aggregated representation that allows to analyze temporal networks from the perspective of spectral graph theory. With this we provide the first analytical explanation for the frequently observed slow-down of information diffusion in empirical non-Markovian temporal networks. We derive an analytical prediction for the magnitude of this slow-down and validate our prediction against two empirical data sets. Counterintuitively, we further show that non-Markovian properties can result in a speed-up of information diffusion that can be related to the spectral properties of the underlying temporal network.