EchoFakeD: improving fake news detection in social media with an efficient deep neural network

Abstract

Modern real-world infrastructure systems consist of a coupled and interdependent subsystems. These system can be modelled as interdependent multilayer networks such as communication networks, power networks, transport networks, etc. These networks are prone to cascade failure of intra-layer and inter-layer links due to overload. For example, in power distribution networks, the flow of electric current beyond their capacities may cause to fail the distribution lines. In multilayer network systems, any damage to the links (intra-layer and inter-layer) may cause cascade failure due to redistribution of loads of failed links to the live links. If it is not controlled, it may damage the entire system. However, optimally redistribution of loads of the failed links to live links can minimize the damage of live links due to overload. In this work, we propose a method to optimally redistribute loads of failed links (intra-layer and inter-layer) to the live links in the interdependent multilayer network in the event of cascade failure. For this purpose, we consider three variants of synthetic and two empirical dataset multilayer networks. Our simulation results reveal that optimal redistribution of loads reduces the number of failed links. Besides, it also reduces the amount of extra load (due to failed links) to be redistributed on the live links. It leads to enhance the number of live links and maintaining the robustness of the entire multilayer networks. © 2021 The authors 2021. Published by Oxford University Press. All rights reserved.