Air Force seeks 'virtual wingman' for intel analysts
Connecting state and local government leaders
The conversational digital assistants would make complex analytics possible for nearly anyone, regardless of technical ability.
The Air Force Research Laboratory wants to develop virtual assistants to help analysts process high-volume, complex intelligence data so they can better uncover and interpret patterns in information related to the nation's increasingly agile and deceptive adversaries.
A broad agency announcement describes the Multi-Source Exploitation Assistant for the Digital Enterprise (MEADE) as a "virtual wingman" that makes complex analytics possible for nearly anyone in intelligence or command and control functions, regardless of technical ability. It would not only answer fact-based questions from existing sources, but also use the conversation to give analysts more information for "connecting more dots" by allowing them to direct, clarify and scope automated analytics.
MEADE has two focus areas. The Real-Time Operator-Driven Gist Exploration and Response, or ROGER, is the conversational question-and-answer system that combines search across intelligence sources, natural language processing, recommendation engines and applied analytics. ROGER answers “lookup” questions directly, focuses analytics and leverages user behavior and cues. It is intended to run in a cloud or distributed computing environment.
The Interactive Analytics and Contextual Fusion focus area is dedicated to finding the best course of action for given situations using contextual fusion of information and anticipatory analytics. The software would use structured narratives or a similar common semantic representation to organize information. It would gather, process, analyze and rank analytical results and be able to identify aspects of products (images, graphs/plots, tables, etc.) that answer the question. IACF would also provide success metrics such as response time, precision, probability of correct association and more.
The $25 million, five-year project will be rolled out in phases, with AFRL creating a systems integration laboratory to support the installation, testing, analysis and maturation of deliverables through its Automated Processing and Exploitation Center.
More information on MEADE is available here.
NEXT STORY: When drones are part of the smart city network