Raytheon has secured a $6M contract from the Defense Advanced Research Projects Agency to build a machine learning tool that will work to develop trust between artificial intelligence systems and human users.
The Competency Aware Machine Learning system will learn various capabilities through a video game-based process and validate best practices on meeting objectives, the company said Monday.
Operators may incorporate mission details into the platform, allowing it to provide recommendations and strategies on certain scenarios. The system will also explain the skills it has learned and the conditions under which they were acquired.
“People need to understand an autonomous system’s skills and limitations to trust it with critical decisions,” said Ilana Heintz, principal investigator for the CAML program at Raytheon’s BBN Technologies business.
DARPA intends to develop a framework that can manage its own behaviors and transform ML tools into “trusted” human partners as part of the program.