On October 3, the Army Research Laboratory posted an industry day announcement for a new broad agency announcement topic on artificial intelligence and machine learning.
The Army Research Laboratory (ARL) Broad Agency Announcement (BAA) for Basic and Applied Scientific Research was recently amended to include a new topic under the Information Sciences Campaign. ARL and the Algorithmic Warfare Cross Functional Team (AWCFT) are hosting an Industry day on October 24, 2017 to discuss this new topic, Artificial Intelligence (AI) and Machine Learning (ML), and invite all businesses engaged in AI/ML to submit their capabilities briefing. This all-day event will allow businesses to engage their counterparts, create synergy, and ask questions.
Industry Day Focus:
The AWCFT is a fast-moving effort launched in April 2017 by the U.S. Deputy Defense Secretary to accelerate the Department of Defense’s integration of big data, artificial intelligence, and machine learning. Also known as Project MAVEN, the AWCFT’s first task involved the development and integration of computer vision algorithms needed to augment and assist military and civilian analysts that are heavily encumbered by the sheer volume of Full-Motion Video (FMV) data the DoD collects.
While the AWCFT’s work against the FMV problem is ongoing, progress has been encouraging, and the AWCFT is now preparing to partner with industry to integrate big data and machine learning against other challenging problem sets across the Defense Intelligence Enterprise. In Project Maven Phase II, the AWCFT will continue to aggressively pursue its objective of turning the enormous volume of data available to the DoD into actionable intelligence and insights at speed.
In order to develop a prototype process, DoD will continue its pursuit and initial application of computer vision algorithms to support Processing, Exploitation and Dissemination (PED), including: computer vision model development focusing on widearea motion imagery, tactical UAV full motion video, and medium altitude full motion video, instantiation of a model development compute cloud with appropriate hardware, data science support, model output database (index and search), AI interface development, and integration support with forward units.
In the latter half of the period of performance, Project MAVEN will expand the prototype from the analysis of vertical image and video to document analysis, natural language processing for machine translation and gisting, optical character recognition, horizontal still photo video object and persona identification, horizontal video object and persona identification, cognitive computing for target systems analysis and entity relationship identification, and other areas using machine learning approaches.
Full information is available here.