Army Developing Shared Toolkit for AI Developers

The Army is working with industry and academia to create a shared toolkit or a “common platform” for artificial intelligence developers. 

The effort is a collaboration among the Army’s AI Task Force, Carnegie Mellon University and machine learning company Visimo, Breaking Defense reported.

Isaac Faber, chief data scientist at AITF, said the toolkit will improve the efficiency of new AI projects by eliminating the need for them to start from the ground up. 

Faber said the Army and its partners achieved a “minimum viable product” or a functioning bare-bones version within the first five months of development. 

He added that the AI development toolkit will help consolidate projects and accelerate capabilities like multi-domain operations and joint all-domain command and control.  

In 2018, the Air Force released a multi-domain operations implementation plan. Multi-domain operations are described as the ability to “seamlessly shift between domains, components and regions to create high velocity, precision warfighting effects to satisfy the Joint Force Commander’s mission needs.”

The Department of Defense more recently described multi-domain operations as the concept of achieving a competitive advantage over an adversary by exposing it to multiple threats that require separate responses.

JAD2C is the Pentagon’s concept of connecting sensors from all of the military services into a single network. The services traditionally developed their own tactical networks that were incompatible with those of the other services. 

Faber said “about a half-dozen” organizations are currently evaluating the AI toolkit, which he said will soon be tested on a sample, unclassified data set.