Large-Scale graph analytics for cyber network vulnerability analysis

Agency: Pacific Northwest National Laboratory (via the Distributed Analytics and Security Institute)
Researchers: Medal, H.R. (PI), Bian, L., Hu, M., Marufuzzaman, M., Zhang, S.
Amount: $600,000

The main goal of this project is to develop a new cyber-node classification approach that utilizes both graph theory and probability modeling. We will pursue our project goal via four objectives: 1) develop an ensemble of models for the probability that a node is malicious or compromised and that an edge exists or not; 2) use advanced model selection techniques to recommend the best probability model; 3) develop novel graph analytic methods that can incorporate uncertainty and scale up to solve very large datasets, and 4) develop new approaches for computing graph uncertainty.