Science Fiction authors perfected visions of thinking machines. Their imaginations were not limited to technical reality. The mind was a topic with insufficient foundational knowledge.

Logic could be modeled into mechanisms. A succession of logical choices could mimic thought. Some thought this promised a road to building a thinking machine. Not TEM Software, we endeavored to build capable assistants to thought.

Tomahawk

In 1985, through a subcontract to Martin Marietta, we became a member of Project Tomahawk. The world’s first high level practical real-time AI autonomous vehicle. The vehicle was developed a decade earlier and was being adapted from the US Air Force to the US Navy. It didn’t begin as a machine learning mechanism. It was taught to fly, navigate, evade enemy radar, and so much more.

LANTIRN

Our work with computer vision included low altitude navigation later incorporated into the known as LANTIRN.

VLS

Our next effort, through Martin Marietta, was the “Vertical Launch System” known as VLS, now used as the primary Navy mechanism to launch missiles for many navies. The user sets the destination and presses a fire button. Behind the scenes is a complex array of sequential robotic processes.

MMIS & FMIS

We went on to improve shop floor throughput by authoring Martin Marietta’s Maintenance Management Information System that included “Just In Time Inventory Control” and automated supply requisitions and ordering. This was followed by our Facilities Management Information System involving sophisticated construction projects. A famous project included the rebuilding of the launch pad at Vandenberg after it was destroyed by a missile exploding upon launch.

The rebuild occurred in record time in part due to the effective use of the Facilities Management Information System.

Hazardous Waste Tracking

In 1990 a subcontract to Westinghouse involved an automated tracking system for hazardous waste from cradle to grave. “Westinghouse’s Environmental Systems” faced E.P.A. fines exceeding $100,000 daily until the product we authored reduced those fines to zero. Again practical AI with a heavy R.P.A. element and a decision matrix advising engine to guide hazardous waste incineration while optimizing landfill selections.

Patent Pendency Modeling

In 2003 a subcontract to Lockheed Martin led to a modeling mechanism for Patent Pendency at the “US Patent and Trademark Office”, followed by a human resource modeling mechanism. These mechanisms contributed to High Level Practical AI Decision Support System to optimize resource planning and accurately forecast pendency measurements. Eventually a 7 year patent pendency was reduced to 16 months. This increase in throughput occurred during a rise of 365,000 patent applications in 2003 to 650,000 applications in 2021.

Pe2e

Also a credit to TEM Software was USPTO’s transformation from manual systems in 2003 to nearly fully computerized systems today – particularly in our efforts support the Patent end-to-end program through our subcontract to Science Application International Corporation. Pe2e modernized many of the legacy systems we contributed to prior to 2012 through our subcontract to Raytheon.

JAIC

Latest efforts include our prime contract with the Joint AI Center of the US Army. During our history we have rolled out a few practical AI products.