Every day, hundreds of chat messages take place between the pilots, the crew and the controllers of the air mobility order 618th Air Operations Center (AOC). These controllers run a fleet of planes at a thousand years old, juggling with variables to determine the tracks to be covered, how long in food or loading will take, or which can pilot these missions. Their mission planning allows the US Air Force to quickly meet national security needs worldwide.
“It takes a lot of work to obtain an anti-missile defense system around the world, for example, and this coordination was carried out by phone and e-mail. Now we use the cat, which creates opportunities for artificial intelligence to improve our workflows,” explains Colonel Joseph Monaco, the director of strategy at the 618TH AOC, which is the largest air operations center in
The 618th AOC sponsors the Lincoln Laboratory to develop these artificial intelligence tools, thanks to a project called Technology for Transition (CAITT) (CAITT).
During a visit to the Lincoln Laboratory of the 618th AOC’s headquarters at Scott Air Force Base in Illinois, Colonel Monaco, Lieutenant-Colonel Tim Heaton and Captain Laura Quitiquit met researchers in the laboratory to discuss Caitt. Caitt is part of a broader effort to transmit AI technology to a major Air Force modernization initiative, called the next generation information technology for improving mobility preparation (NitMre).
The type of AI used in this project is the treatment of natural language (NLP), which allows models to read and treat human language. “We use NLPs to map the main trends in cat conversations, recover and cite specific information, and identify and contextualize critical decision points,” explains Courtland Vandam, researcher at Lincoln Laboratory AI AI technology and systems groupwhich leads the project. Caitt encompasses a series of tools taking advantage of NLP.
One of the most mature tools, the summary of subjects, extract from the trend subjects of chat messages and formats these subjects in a friendly display highlighting critical conversations and emerging problems. For example, a subject of trend can be read: “the members of the crew missing from Congo visas, potential delay”. The entry shows the number of cats linked to the subject and summarizes in Bullet Points the main points of conversation, based on specific cat exchanges.
“Our missions depend on time, so we must quickly synthesize a lot of information. This feature can really tell us where our efforts should be concentrated,” explains Monaco.
Another production tool is semantic research. This tool improves the search engine of the cat service, which currently returns empty results if chat messages do not contain each word of the request. Using the new tool, users can ask questions in a natural language format, such as why a specific plane is delayed and receive smart results. “He incorporates a research model based on neural networks that can understand the intention of the user of the request and go beyond the correspondence of the term,” explains Vandam.
Other tools under development automatically aim to add users to chat conversations deemed relevant for their expertise, to predict the amount of field time necessary to unload specific types of cargoes from aircraft and to summarize the key processes of regulatory documents as a guide of operators as they develop mission plans.
The Caitt project was born from the DAF-MIT AI accelerator, a three-part effort between MIT, Lincoln Laboratory and the Air Force department (DAF) to develop and transition algorithms and AI systems to advance DAF and society. “Thanks to our involvement in the AC accelerator via the Nitmre project, we realized that we could do something innovative with all unstructured cat information in the 618th AOC,” explains Heaton.
While laboratory researchers are advancing their prototypes of Caitt tools, they started to get them to 402nd Software Engineering Group, a software supplier for the Ministry of Defense. This group will implement the tools in the operational software environment used by the 618TH AOC.