Automated Tools for Adversarial Threat Characterization

ABSTRACT: The objective of the proposed Space Situation Awareness via Predictive Intelligent Reification of multi-INT Threat-models (SPIRIT) effort is to develop a real-time, multi-INT SSA fusion capability that provides predictive insight into both deliberate and environmental threats facing space assets. A key enabler of this capability is the application of semantic web technology towards the enrichment and fusion of multi-source observations. In Phase I of this SBIR the ISS team successfully demonstrated several powerful capabilities including: a capability to extract, semantically tag, and temporally align knowledge from unstructured data sources; how to extend semantic web technologies with temporal and probabilistic reasoning in support of multi-INT fusion; and how learned multi-INT indicator models can be applied towards predictive SSA. The primary objective of Phase II of this effort are to mature the functional prototypes developed under Phase I into a comprehensive JDL level 2/3 fusion service capable of supporting real-time, multi-INT predictive SSA within the JSPOC Mission System (JMS). In addition to a significant software engineering effort, SPIRIT Phase II also includes significant research into several challenging problems including semantic based fuzzy membership functions, multi-INT entity disambiguation, and assigning confidence to knowledge extracted from unstructured text. BENEFIT: SPIRIT will provide an extensible framework for supporting SSA in a net-centric environment that will be compatible with the new Joint Space Operations Center Mission System (JMS) SOA architecture. Using a federation of normalcy, threat, and COA models, the real-time, multi-INT SPIRIT system will be able to rapidly identify and characterize threats, be they intentional or natural. Consuming streaming multi-INT data as well as open source text, the semantic enrichment capabilities of SPIRIT along with its multi-logic reasoning system will support predictive awareness of emerging threats from partial evidence, enabling timely threat identification and development of effective countermeasures. Although initially developed to support SSA, the declarative knowledge framework can be adapted to support new domains with minimal if any software programming. The initial commercial applications of SPIRIT include SSA and DCS systems such as JMS.