Data to Decisions Applied Research
Abstract
The Joint Data Management Program has been restructured in FY 2012 to become an expanded Data-to-Decision program. This Data-to-Decision program builds on the FY 2010 and FY 2011 accomplishments with increased objectives and technology developments critical to on-going operations. That program had two subtasks as outlined below: Data Shaping for Exploitation - When tracing the information processing chain from the sensor inputs to the user/analysts, the automated techniques that are known and can be applied become fewer and less mature. The simple information processing chain goes from (1) data tagging and (2) pre-processing to (3) multi-source common data representation to (4) triage/identify high priority data subsets for analysis and action. Candidate research topics to be explored include pattern analysis, data classification for important and prioritization, criticality assessment, change detection, uncertainty management and reduction, high level structures, data search and retrieval, feature extraction, automatic translation, and automated or assisted pattern recognition. Data Discovery for Exploitation - In order to better discover and exploit the growing amount of sensor data, the following areas of research are considered: object recognition in scenes and streams, discovery and exploitation at the edge, structuring knowledge for discovery, improving analytic throughput, aidingIntelligence, Surveillance and Reconnaissance (ISR) functions, layered analysis and interpretation, effects prediction for decision support, and cross domain access for effective ISR. These two tasks will be consumed within a new structure in the Data-to-Decisions program. This new program will focus on developing open-architecture technologies for decision support systems to help reduce future development time and cost of data management, analytics and user interface subsystems. The program will use a spiral development model with four-steps. Each year Operational teams will choose a series of cross-service challenge problems dominated by a specific sensing modality. Representative data for each of those problems will then be collected for testing against that problem. A Development team will design algorithms and data management architectures using high-level languages and self test on controlled data sets to address those challenge problems. Independent assessment will occur with sequestered data sets, but each development tool will also be tested against new sensors not included in the self-testing to determine fragility. A Transition team will host the developed algorithms as services in a spiraling prototype system. The Applied Research program will concentrate on the Development portion of this collaborative effort, while the Advanced Technology Development program focuses on the infrastructure piece, to include the Operational, Assessment and Transition portions. There will be four thrust areas in total, MOVing INTelligence (MOVINT) analytics, Text analytics, IMagery INTelligence (IMINT) analytics and Integration. Work being conducted under the previous two tasks will be consumed within these new tasks and are appropriately described as below.
Document Details
- Document Type
- Project
- Publication Date
- Oct 01, 2012
- Source ID
- P266_0602663D8Z_2_0400_PB_2012
Related Documents
- Root: Data to Decisions Applied Research
- Child Accomplishment: MOVINT Analytics
- Child Accomplishment: Text Analytics