Department of Computer Science
University of Notre Dame
154 Hurley Hall
Novel Strategies For Efficient Network Mining: Implications For Aging and Disease
Networks (or graphs) have been useful models for many real-world phenomena in various research domains. Examples include technological networks such as the Internet, information networks such as the World Wide Web, social networks such as Facebook, ecological networks such as food webs, or biological networks such as protein-protein interaction networks. Owing to the exponential growth of real-world network data, the complexity of the networks has become the central issue in their modeling and understanding. Hence, there is a need for sophisticated mathematical and computational techniques for mining the networks.
We will present our recent computational approaches for comparing (i.e., aligning) large real-world networks, as well as for studying their evolution, in order to enable efficient extraction of functional information from network structure. Then, we will discuss how the approaches can be applied to biological networks, e.g., networks of interactions between proteins in the cell, to address many important problems in biomedicine, such as identifying novel aging-related genes and thus drug targets. In addition, we will show that the same methods can be successfully used to study other network types as well, such as social networks – networks of friendships and acquaintances or on-line social communities. For example, we will illustrate the application of network analysis to studying how our friendship networks shape our lives, as well as to de-anonymizing on-line social network data, thus impacting user privacy.