Air Force Deep Learning (AFDL)

A Virtual Reality for Multi-Int Deep Learning modeling and simulation environment is proposed which draws from proven advances in modeling and simulation, Big Data technologies and industry leading machine intelligence platforms. RF Physics-based models and simulators focus the environment in the RF domain. A design model emphasizes Big Data control, archive, management, and analytics maintaining traceability of data from generation to Machine Learning/Deep Learning networks. The environment, named Air Force Deep Learning, is an enterprise environment agnostic to domain and machine intelligence networks. Commercial RF Physics-based models and data make the environment specific to the RF domain. The model based design is essential for active RF systems where data provenance from sensor to targets to processing is an imperative for successful machine intelligence applications. The enterprise approach to Air Force Deep Learning offers significant contribution to speed advancement of autonomous system performance. Model based design innovation includes convergence of Model Based Systems Engineering and advanced simulation formalisms necessary for Phase II where more complex and sophisticated systems across Multi-Int domains are considered. The output of the Phase I is a design model and demonstration that the VR-MDL environment output compares with actual data or truth from a representative radar system.