The US Navy is developing the Navy Tactical Cloud Reference Implementation (TCRI) to enable on-demand access to shared sensors and computing resources. They seek to leverage TCRI to perform advanced analytics and reasoning, but this poses multiple challenges. These include addressing significant uncertainty in warfighting domains, the volume of data available through modern sensors and networks, and the time and cost constraints in creating unique analytics for specific data sources and data relationships. To address these challenges, Navalytics will provide a foundation that supports probabilistic reasoning; handles missing, unreliable, and stale data; and supports fusion of multiple diverse input streams.
Navalytics combines Charles River’s Figaro probabilistic programming language with a modular, reusable framework for developing reusable analytics modules. Charles River will add distributed, big data-enabled algorithms to Figaro as part of the Navalytics effort. The reusable architecture for distributed analytics can be used in many domains, such as real-time health monitoring, threat assessment, and intelligence analysis.