Every day, hundreds of chat messages flow between pilots, crew and air mobility commanders 618. Air operation center (AOC). These controllers manage a thousand all the fleet of aircraft, juggling the variables to determine which routes fly, how much time it will take to bake or charge inventory or who can fly these missions. Their planning of the mission allows the American Air Force quick response to the needs of national security around the world.
“For example, obtaining a missile defense system takes a lot of work, and this coordination was made by phone IE -Mail. Now we use a chat that creates opportunities for artificial intelligence to increase our work department work center.
618. AOC sponsors Lincoln Laboratory to develop these tools of artificial intelligence through the project entitled AI technology for transitional (Caitt).
During a visit to Lincoln, laboratory from the headquarters of the main 618. AOC at Scott Air Force Base in Illinois, Colonel Monaco, Colonel Tim Heaton and captain Laura Quitiquit met with laboratory researchers to discuss Caitt. Caitt is part of a broader effort to transform AI technology into the main initiative of the modernization of the Air Force, called the new generation IT technology for improving readiness for mobility (Nitmre).
The type of AI used in this project is natural language processing (NLP), which allows models to read and process human language. “We employ the NLP to map the main trends in conversations at chat, downloading and quoting detailed information as well as identifying and context of critical decision points,” says Courtland Vandam, researcher at Lincoln Laboratory’s AI Technology and Systems Groupwhich conducts the project. Caitt includes a package of tools using NLP.
One of the most mature tools, summary of the topic, distinguishes trends from chat messages and formats these topics in a user -friendly display emphasizing critical conversations and emerging problems. For example, a popular topic may be: “crew members are lacking in Congo’s visa, delay potential.” The entry shows the number of chats related to the topic and summarizes the bullet points of the main conversation points, combining with specific chat exchanges.
“Our missions are very dependent on time, so we must quickly synthesize a lot of information. This function can really indicate where our efforts should be focused,” says Monaco.
Semantic search is another tool in production. This tool improves the chat search engine, which currently returns empty results if chat messages do not contain every word in the question. By using the new tool, users can ask questions in natural language format, such as why a specific aircraft is delayed and receive intelligent results. “Contains a search model based on neural networks, which can understand the user’s intention and go beyond matching the deadline,” says Vandam.
Other developed tools are designed to automatically add users to chat conversations, which are considered important for their specialist knowledge, predict the amount of time of land needed to discharge specific types of load from aircraft and summarize key processes from regulatory documents as a guide for operators during the development of mission plans.
The Caitt project has grown from the AI DAF-Mit AI accelerator, a three-person effort between myth, Lincoln Laboratory and the Air Force Department (DAF) in order to develop and AI transition systems to develop both DAF and society. “Thanks to our involvement in AI Accelerator through the Nitmre project, we realized that we can do something inventive with all unstructured chat information in 618. AOC,” says Heaton.
When laboratory researchers develop their prototypes of Caitt tools, they began to transfer them to 402nd Software Engineering Group, software suppliers for the Defense Department. This group implements the tools for the operating environment of the software used by 618. AOC.