The Defense Advanced Research Projects Agency has started the process of soliciting research proposals for a machine learning technology to help autonomous systems determine its task competency and support human-machine teaming activities.
DARPA said Tuesday in a FedBizOpps notice it aims to develop a framework to transform ML tools into "trusted" human partners.
The agency seeks a new approach for machines to manage their own behaviors and allow users to ontain insight into a system's capability through time-critical and dynamic environments. Additionally, the agency seeks competencies in the robotic navigation, action planning, decisionmaking and object recognition areas.
The project will run for 48 months consisting of a 36-month research period and a 12-month technological demonstration phase.
Interested vendors have until March 6 to submit one-page profiles that indicate their contact information, technical capabilities and preferred expertise from other teams.
Full proposals are due April 22.