Mixed-Initiative Discourse Analysis System (MIDAS)
Department of Defense
Agency Tracking Number:
Solicitation Topic Code:
Small Business Information
Charles River Analytics Inc.
625 Mount Auburn Street, Cambridge, MA, -
Socially and Economically Disadvantaged:
AbstractABSTRACT: Intelligence analysts across the Department of Defense identify and act on threats to national security. Their analyses must encompass both individuals and groups of potential interest, which analysts can identify by examining discourse in communications made by individuals through blogs, message boards, and other social media. Specific patterns in this discourse inform analysts"understanding of emerging groups and their intent, although purely automated techniques separate the analyst from the nuanced information that is critical to effective understanding of human behavior. We propose to design a mixed-initiative approach that combines automated techniques with user-centered tools to allow analysts to: (1) express, explore, and share their own reasoning about emerging groups and how individuals align themselves with groups; (2) proactively identify groups and their intent; and (3) direct collection assets to confirm or deny analysts"assertions. Such a capability will enable analysts to make more effective and efficient decisions about group behavior by combining automated discourse analysis with analyst-directed reasoning, and allow analysts to respond more rapidly and cost-effectively to current and emerging threats to national security. BENEFIT: The anticipated benefits include the advancement of discourse analysis methods that map to social psychology features, and the use of these features to support analyst-centered reasoning in a mixed-initiative tool. Military and commercial applications of this effort include transition to the intelligence community, as well as elements of NASIC, AFTC, and MISO. Military applications include anticipatory ISR, COA analysis, influence operations, and stability and reconstruction operations. In addition, we plan to transition specific causal influence modeling features of the MIDAS system to our BNet software suite to increase its commercial viability.
* information listed above is at the time of submission.