Santo Fortunato (left) and Filippo Radicchi
Professor Santo Fortunato and Associate Professor Filippo Radicchi of the Luddy School of Informatics, Computing, and Engineering have been awarded a grant worth nearly $450,000 from the Army Research Office to study the multilayer networks and their application to real-world problems using machine learning and artificial intelligence methods.
Interest in applying machine learning methods to the analysis of networks is growing. Most of the research focuses on analysis of isolated graphs, but Fortunato and Radicchi maintain that since networks in the real world are coupled with other networks, analyzing multilayer networks could highlight the presence of multiple layers of interactions.
“For example, in social media, the same person may interact with different persons depending on the online platform,” Fortunato said. “Studying how opinions are shaped by those social networks requires us to fully account for potential coupling across platforms; studying each platform in isolation does not provide a realistic description of the actual process of opinion formation.”
Although the project will be mostly theoretical in nature, Fortunato and Radicchi will explore other applications as well, including network robustness, identification of influencers, and navigability.
“Understanding what features make a multilayer network robust is very important in designing critical infrastructures able to maintain a functional state even in scenarios of extreme stress or capable to quickly recover after systemic collapses,” Radicchi said. “We aim to create novel methods for identifying the features that a robust network should have. We expect to improve the state of the art by proposing some innovative measures that can help to spot influencers. The navigability problem consists in providing efficient exploration strategies of a city using multiple transportation layers, e.g., subway and bus.”
The researchers will initially test the ability of existing embedding techniques to reproduce key structural features of networks with a focus on community structure. Learning the limits of existing techniques will provide valuable input on new embeddings that can overcome such limitations.
“For me, this project is a big opportunity to delve into the increasingly stronger connection between artificial intelligence and networks,” Fortunato said. “This project will allow me to deepen my knowledge of artificial intelligence and to see how it can help us solve open problems regarding networks.”
The Army Research Office was founded in 1951 and drives cutting-edge and disruptive scientific discoveries that will enable crucial future Army technologies and capabilities through high-risk, high pay-off research opportunities. It aims to increase fundamental knowledge and understanding in the chemical, life, physical, engineering, materials, mechanical, computing, information, network, mathematical, earth, and social sciences, related to long-term national security needs.
“The award represents a natural continuation of two of my past awards—the National Science Foundation CAREER award and another ARO project—focusing on the theoretical analysis of multilayer networks,” Radicchi said. “The current project will take advantage of machine learning methods, and the focus will be directed towards the understanding of dynamical processes occurring on multilayer networks.”
The ARO grant provides funding for the research over the next three years.
“Analyzing the relationship between networks can provide so much insight that can help solve real-world problems, which is one of the main goals of our school,” said Kay Connelly, the associate dean for research at the Luddy School. “The work being conducted by Santo and Filippo will be enlightening and offer opportunities for more innovation in the future.”