Challenges in community discovery on temporal networks

Community discovery is one of the most studied problems in network science. In recent years, many works have focused on discovering communities in temporal networks, thus identifying dynamic communities. Interestingly, dynamic communities are not mere sequences of static ones; new challenges arise from their dynamic nature. Despite the large number of algorithms introduced in the literature, some of these challenges have been overlooked or little studied until recently. In this chapter, we will discuss some of these challenges and recent propositions to tackle them. We will, among other topics, discuss of community events in gradually evolving networks, on the notion of identity through change and the ship of Theseus paradox, on dynamic communities in different types of networks including link streams, on the smoothness of dynamic communities, and on the different types of complexity of algorithms for their discovery. We will also list available tools and libraries adapted to work with this problem.