Multi-Level Associative Content Environment (MACE)

Intelligence successes regularly require analysts to quickly wade through massive stores of uncertain data to associate and correlate concepts, individuals, locations, and resources. Current data systems support massive data search and retrieval, or automated analysis, but lack the flexibility to do both well. What is needed is a system that can balance between these two, to maintain and rapidly navigate association data at multiple levels of detail, while avoiding information loss that occurs when too much or too little data is considered. This support must happen at operational tempo, correlating new data streams with historical archives to detect and track threats and develop situational awareness.Accordingly, we are developing MACE, a multi-modal association database management and analysis system implemented as a multi-level graph. In Phase I, we built a system design, and conducted a proof-of-concept demonstration to show that MACE will scale to operational data volumes. In Phase II, we will extend MACE to additional modes of data, improve entity resolution, and provide recommendations to ensure robust analysis. Using graph databases, we will achieve analytic and run-time performance successes where traditional databases fail. MACE will leverage existing open software in a plug-and-play architecture to provide an open, license-free solution.