Social Network Modeling and Intent Recognition

NSI is developing a set of automated tools for terror threat network analysis, tracking and intent/capability recognition. This included identifying and implementing techniques for stochastic modeling of terrorist social networks and their intentions. The approach taken included: probabilistic social network tracking on tagged, aggregated data sets across multiple relations; red team scenario modeling including a new terror attack description language; scenario-driven intent recognition via dynamic Bayesian networks; and combined use of external metadata and content-based metadata. As part of this work, we are exploring modeling and simulation (M&S) tools for scenario and corpora library creation, quantifying the tradeoffs between gaming, training, and M&S approaches. We are also designing and implementing a standalone attack scenario generator tool to create terror attack scenarios as part of a larger testbed.