Exploiting Essential Elements of Information from Significant Activity Reports (SIGACTS) for Forensic Analysis

Award Information
Department of Defense
Air Force
Award Year:
Phase I
Agency Tracking Number:
Solicitation Year:
Solicitation Topic Code:
Solicitation Number:
Small Business Information
1892 Mill Run Court, Hellertown, PA, 18055
Hubzone Owned:
Minority Owned:
Woman Owned:
Principal Investigator:
William Pottenger
Principal Investigator
(484) 851-3423
Business Contact:
William Pottenger
(484) 851-3423
Research Institution:
There has been a great deal of development effort within the academic, governmental, and commercial arenas over the past few years in the field of information extraction. Based on our investigation, however, it appears that there is no ground truth for SIGACTS reports, which prevents GMTI users, program managers and researchers from performing apples-to-apples comparisons of information extraction technologies. Given our close working relationship with GMTI analysts and experience with GMTI data in an ongoing project, we propose to address this deficiency as part of this SBIR project. Another critical issue is the need to link information extracted from SIGACTS with other types of information used by analysts working with GMTI data. We have been working closely with GMTI analysts at AF ESC CEIF to apply data analytics technologies to pattern discovery in GMTI data, and are well positioned to link information extracted from SIGACTS reports to GMTI data. This is another goal of this SBIR project. We also have recently developed a semi-supervised active learning method for information extraction that can be leveraged in existing systems for information extraction such as Intuidexfs IxEEEf> or Janyafs Semantexf>. A third goal of this SBIR project is thus to apply this rule learning technique to SIGACTS reports. Finally,it is also known that information extraction techniques can be supplemented with text classification techniques. In our work we have developed an approach to text classification that leverages recent advances in statistical relational learning. The final thrust of this SBIR will be to apply our text classification algorithms to the problem of automatically categorizing events extracted from SIGACTS reports. BENEFIT: The primary customers for SIGACT extraction products are forensic GMTI analysts at JIEDDO/COIC, NASIC, and NGA that will benefit from overlaying SIGACT information over GMTI data. There are also near-real-time analysts in the Army, Marines and on board JSTARS and LSRS platforms. Even though NRT analysts do not perform in-depth analysis of forensic data, their ability to interpret NRT data will be enhanced by SIGACT data. According to our discussions, analysts have a strong preference for new capabilities to be integrated into the products they currently use instead of developing new tools. Currently, analysts use the MOVINT Client (MC) (developed by Northrop Grumman under funding from AFRL/Rome fusion lab), ISR Forensics (MITRE) display and Google Earth-based products. Army analysts are also using Common Ground Station (CGS). Most analysts are scheduled to transition to the new DCGS (Army or AF versions) that will include MC. In addition, ground station for Army VADER UAV (developed by BAE Advanced Information Technologies) will also be based on MC. We are currently in discussions with the AFRL Fusion lab to develop an open API necessary to integrate external products such as the extraction products proposed herein into MC.

* information listed above is at the time of submission.

Agency Micro-sites

SBA logo

Department of Agriculture logo

Department of Commerce logo

Department of Defense logo

Department of Education logo

Department of Energy logo

Department of Health and Human Services logo

Department of Homeland Security logo

Department of Transportation logo

Enviromental Protection Agency logo

National Aeronautics and Space Administration logo

National Science Foundation logo
US Flag An Official Website of the United States Government