All-Source Precision Natural Language Processing (ASPN) for Intel Analysis and Exploitation

In today’s Information Environment (IE), the increasing volume, velocity, variety, and value of data presented in a myriad of structured, semi-structured, and unstructured formats/reports can overwhelm intelligence analysts and decision makers across DCGS and related programs. This increasing volume of information degrades the analysts’ ability to process, exploit and disseminate time-critical, mission-sensitive analysis as report traffic is manually reviewed, correlated, and validated on a continuous basis. To address this challenge, the inclusion of a semi-automated to fully-automated Natural Language Processing (NLP) analysis and exploitation capability is needed to augment and complement intelligence analysts. The existing suite of applications (e.g. Analyst Notebook, ArcGIS) to support the intelligence analyst do not address the rapidly growing critical responsibility of extracting knowledge from unstructured text documents. More specifically, the accuracy and completeness required by the knowledge extraction needs of intelligence analysts are not supported due to technological and design limitations of current market offerings. Polysentry ASPN is will deliver on the key mission to increase NLP analysis accuracy and completeness within a scalable and standards-based system architecture that enhances the Intelligence Analyst’s ability to process massive quantities of daily report traffic or other forms of documents. Both false positives (e.g., undesirable noise due to incorrect interpretations) and false negatives (e.g., missed linkages and correlations) generated by existing artificial intelligence/machine learning (AI/ML) applications diminish analysts’ ability to understand and exploit Intelligence, Surveillance, Reconnaissance (ISR) information holistically and at the speed of relevance. To address the above, Polysentry proposes to adapt Polysentry’s All-Source Precision NLP (ASPN) platform, a commercial product currently supporting large enterprise customers, to integrate with DCGS program's hardware and software configurations. Core components of the existing ASPN system’s engineering are adaptability and extensibility for easy deployment to any customer’s specific operational needs and technical requirements. Built on Polysentry’s commercially successful ASPN NLP platform, ASPN–MC will perform Knowledge Extraction capabilities to automatically identify geospatial, temporal, and contextual connections contained within identified data sources (including structured, semi-structured, and unstructured reports), quickly correlating intelligence in a fully interactive, human-driven, machine-assisted analysis experience.