The Air Force is researching advanced automation systems to combat threats worldwide. Now it wants an automated system to help combat bureaucracy at home.
The service is hoping to develop a supercomputer, a “cognitive thinking machine,” that can analyze vast quantities of data, track cost and benefits, and easily navigate the library of U.S. government codes and regulations, all in an effort to improve the acquisitions process.
“There are thousands of pages of policies, laws, and regulations that affect Air Force acquisitions,” Camron Gorguinpour, director of transformational innovation for the Air Force, said in a statement.
“We need to create a baseline platform and teach the system how to understand context so that it can answer questions accurately and become a resource that personnel can access,” he said. “Of course this is an initial effort; however, over time I expect these types of tools can help people in the Air Force, government and industry better navigate what is a very complex bureaucracy.”
The goal is to create a system that will help military officials and contractors — especially small businesses — navigate the complex world of Defense Department acquisitions.
The Air Force pointed to a 2006 study by the Government Accountability Office, Congress’ top watchdog, which found that “the challenge of operating in accordance with complex federal acquisition regulations discourages small and innovative businesses from partnering with the government in emerging markets.”
The Air Force hopes the new system will be an advanced tool, making it easier for businesses to understand the requirements of a contract and to get any of their questions answered immediately by the computer system.
Applied Research in Acoustics, in Washington, D.C., and KalScott Engineering, in Lawrence, Kansas, won the initial contract in July to build a “natural language query system … to provide users insights into defense contracting statutes, regulations, practices, and policies,” the Air Force said.
The natural language would help users interact more easily with the program, said Suman Saripalli, the owner of KalScott Engineering.
Computers give pre-programmed or formulaic answers, but the project would hopefully create a computer that could more easily present information "how a person would," Saripalli said.
"Basically the Air Force is looking for a way to make these searches easy and intuitive for people," he said. "You would ask the system a natural question and it would come back with a natural answer."
"This would be an application that would allow users, and they could be us in the business community as well as on the government side, to quickly search and mine all this data that’s available in the files," Saripalli said.
Jason Summers, chief scientist at ARiA, said in a statement that "automation needs to be an enhancement to the experience of a user in performing her job, not a hindrance or a complete substitute for human judgment."
"Automation is only useful when it earns the trust of the user," he said. "To do that, it has to do two things. It has to be correct and it has to have a reasonable justification for its judgments."
The initial program should be up and running by 2018.