From a static to a dynamic analysis of complex networks

Physical contacts between individuals, social interactions, economic transactions, or computers exchanging packets, all these different systems have in common to be a set of interacting objects without being coordinated by a central brain. Therefore, the structure of interactions results from decentralized processes, which are often unknown. Since the 90s, it has been pointed out that graph representations of such systems exhibited common properties, allowing to use transversal methods to describe them and understand their underlying mechanisms. These studies then evolved into a unified field of research, which is called complex networks analysis.
Because of the simplicity of graph representations, as well as the rich body of knowledge accumulated in graph theory and algorithmics, describing interaction data with graphs has led to substantial successes. However, increasing access to online datasets highlighted the need to take into account the intrinsically dynamic aspect of interaction data. My research work tackles several aspects of the evolution from a static to a dynamic description.
It is organized around three main axis: the first one deals with the description of dynamic processes on evolving networks, and more specifically spreading phenomena. The second axis concerns the issue of recovering and predicting interactions in a temporal network. The third one engages in the question of modeling the interaction structure by generating random networks that mimic real data.

2 Doktoranten (Forschungsleitung (Direction de recherche) / Co-Betreuung)

• ARHACHOUI Nouamane : Mitigation strategies against Fake News diffusion in online social platforms
• CASSELLS Duncan : Polarization in large socio-informational networks: models and measures

6 Doktoren 2016 - 2023