Multilayer network embeddings and applications to real-world problems
Abstract
Network science has created a wealth of tools and models that can be applied to the analysis of many different systems. But many networks do not live in isolation, instead they interact with one another, as layers of a network of networks. Such multilayer networks are everywhere: from infrastructure networks formed by systems such as the power grid, the Internet and financial markets, to multilayer social and communication networks, to the brain and biological networks of the cell. As we cannot understand the living cell if we do not integrate information from all of its biological networks, for many other systems the function of one network is often interdependent on the function of another network. This is why multilayer networks have received a lot of attention over the past decade. Recently, many techniques have been developed to create representations (embeddings) of networks in geometric spaces, while preserving their properties. The task is performed with the goal of reducing the complexity of the network and to exploit the much greater flexibility of operating on the simple geometric structures that ensue. Indeed, it turned out that embedding techniques allowed researchers to find better solutions to standard tasks like node classification, link prediction and network visualization than established network techniques. In this project novel techniques to project multilayer networks onto space will be developed, by leveraging networksÕ community structure, so that points representing nodes belonging to the same community will be located near each other. The potential of these geometric embedding techniques for networks of networks will be assessed by applying them to three specific challenges: robustness optimization, navigability and detection of influencers. Robustness is defined as the ability of the system to remain functional as it experiences failures of its individual elements. Understanding what features make a network of networks 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. Multilayer network embeddings allow for the creation of novel methods for the identification of the features that a robust network should have, by focusing on the geometric similarity of the layers. The navigability problem consists in providing efficient exploration strategies of a geographic area (e.g. a city) using multiple transportation layers, e.g., train, subway and bus. A related issue is the interplay between the structure of the network and the navigation strategy. Efficient navigation algorithms traditionally designed for graphs, like greedy search protocols, will be adapted to the geometric representations provided by the embeddings. Identifying the most influential spreaders in a network is a significant step towards improving the use of existing resources to speed up the spread of information for application such as viral marketing or hindering the spread of information for application like virus blocking and rumor restraint. Identifying influential spreaders on a communication network where users interact via different platforms is a highly non-trivial task, due to the complex way that a user can affect the spreading process by acting on different media. Network embeddings will be used to define non-conventional centrality metrics useful for the identification of optimal spreaders. The educational goal of the project is to integrate specific courses dedicated to the analysis of networks of networks into the Ph.D. and Master programs in Complex Systems and Engineering at the Indiana University Luddy School of Informatics, Computing, and Engineering. Also, through partnerships with Historically Black Colleges and Universities, the outreach plan will recruit students from minority and other under-represented groups.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 25, 2021
- Source ID
- W911NF2110194
Entities
People
- Santo Fortunato
Organizations
- Army Contracting Command
- Indiana University
- United States Army