Traditional risk-assessment methods are often insufficient in these data-poor environments. The project focused on using graphical models to identify high-value targets, estimate attack likelihoods, and compute actionable Annualized Loss Expectancy (ALE) with quantified uncertainty.

The students identified the most critical 1% of network assets whose compromise could destroy the entire system, flagged high-risk connections and potential targets for attackers, and simulated the cascading spread of compromises across interconnected networks.