Mirkwood is a UT Starters grant
project summary
Quantitative analysis of risk models, such as fault trees and attack trees, is essential for ensuring the reliability of high-tech systems. However, existing methods assume the system owner has access to complete, objective data. This condition is often not met in practice, especially in cybersecurity, where complete data can be hard or impossible to obtain.
MIRKWOOD's aim is to develop new, fast algorithms for quantitative risk model analysis under data uncertainty. To do this, we will combine three exciting new developments in the field of formal risk analysis:
- Algorithms rooted in polynomial algebra, that provide computational efficiency;
- Uncertainty frameworks such as Dempster-Shafer theory and fuzzy numbers, that provide the ability to handle incomplete data;
- The category-theoretic viewpoint of operads, that provides the means to generalize our results to many risk models, risk metrics, and uncertainty frameworks.
researchers
