Innovative approaches to Situation Modeling, Threat Modeling and Threat Prediction

The technical objective is to improve High-Level Information Fusion (HLIF) robustness by adaptive use of external"data repurposing". The approach provides the functional decomposition and problem-to-solution-space mappings by extending the Dual Node Network (DNN) Data Fusion & Resource Management (DF & RM) Technical Architecture at Level-4 for repurposed data pattern discovery and HLIF context assessment and conformity management (CACM). The DNN integrates across DF & RM levels, permits reuse of designs and software, and prevents one-of-a-kind solutions. The DF & NN team (including world-class fusion experts) has developed affordable methods to discover unknown models without truth data that address bias and uncertainty-in-the-uncertainty issues for'big data'having non-numeric qualitative reports. The semi-supervised methods enable operators to construct repurposed data models mapped into HLIF ontologies based upon a small subset of data. The data-driven methods automatically learn to characterize such repurposed data and learn the correlations with HLIF products which are used to automatically find, characterize, track, and show relevant context for abnormal non-conforming repurposed database behaviors. We will be developing and testing the repurposed data machine learning and CACM prototype on either the real GPS-related Signal-to-Noise-Ratio & space weather for JSpOC Mission System (JMS) or the SYNCOIN data for Distributed Common Ground Station (DCGS) transition.