Seminario de investigación | Optimization Strategies for the Inverse Eigenvector Centrality Problem: An Application to the Political Blogosphere

Seminario de investigación

Optimization Strategies for the Inverse Eigenvector Centrality Problem: An Application to the Political Blogosphere

Expositor | Fabio Raciti | University of Catania

The study of complex networks provides fundamental tools for understanding interactions across diverse systems, from infrastructure to social dynamics. Central to this field are centrality measures, which quantify the influence of nodes within a given topology. Traditionally, research focuses on the forward problem: calculating centrality based on a fixed adjacency matrix. However, many modern applications demand the Inverse Centrality Problem: modifying edge weights so that the network yields a pre-defined centrality vector.
In this talk, we explore various optimization criteria for the Inverse Eigenvector Centrality Problem, focusing on how different reweightings can achieve global influence targets. Finally, we apply our methodology to a classic real-world case: the network of US political blogs during the 2004 presidential election. We demonstrate how our optimization strategies perform in a deeply polarized environment, offering insights into the robustness and manipulability of political discourse.

Sobre el expositor

Fabio Raciti is an Associate Professor of Operations Research at the University of Catania, Italy, and has obtained the Italian National Scientific Qualification for Full Professor in Operations Research.

His main research interests include inverse problems, Nash equilibrium problems, variational inequalities in probability spaces, network games, and optimization methods for network centrality measures.

He is the author or co-author of more than 75 research papers and of the research monograph Uncertainty Quantification in Variational Inequalities: Theory, Numerics, and Applications, published by CRC Press in 2021.