Automated Identification and Semi-automated Resolution of Conflicting, Suspicious, Inconsistent and Deceptive Information

ABSTRACT: This proposal details a software suite which will be developed for the automated identification and semi-automated resolution of conflicting, suspicious, inconsistent and deceptive information (CSI-Info). The methods employed will adapt previous research work performed by the PI"s while advancing the work to better meet the needs of the challenging CSI-Info identification and resolution domain. The multi-stage software suite will include processing elements to: (1) Recognizing and resolve formatting inconsistencies (specifically data field formats); (2) Perform common referencing and uncertainty alignment (unit conversion and qualitative/quantitative observation transformation and comparison); (3) Perform anomaly detection for statistical and rule-based anomalies; (4) Identify data association transitivity requirement violations (to make sure that if X is similar to Y and Z, then Y and Z should also be similar for consistency sake); (5) Extend stochastic graph matching technique to identify exclusive OR (XOR) violations; (6) Develop human touch points for the resolution of CSI-Info (for the cases where resolution was not automatically attained). These research and implementation tasks will be performed over a period of 9 months with intermediate system testing and evaluation summarized in a deliverable format. BENEFIT: The research study and software prototype will provide an analyst with a range of CSI-Info identification and resolution mechanisms. The application of these methods can improve entity resolution, data association, common referencing, and ultimately fusion and situation understanding activities. These in turn can lead to thwarting insurgencies, uncovering terror cells, preventing IED attacks and result in stable and peaceful nation states. In commercial applications these technologies can help avert financial crime, credit card fraud, and result in fewer loan defaults, for example. The improved decision-making resulting from CSI-Info detection and resolution can save many lives and save millions in financial wealth.