Jamie Bennet
The Defense Advanced Research Projects Agency selected four teams for a new program aimed at developing algorithms and architectures as the foundation for trustworthy artificial intelligence systems.
The Assured Neuro Symbolic Learning and Reasoning program will focus on combining data-driven machine learning and traditional symbolic reasoning to build robust and reliable autonomous systems, DARPA said Monday.
ANSR is a three-phase initiative that will begin with gaming environment-based design and end with a live demonstration of autonomous intelligence, surveillance and reconnaissance prototypes for the Department of Defense.
The University of California, Berkeley, Vanderbilt University, Collins Aerospace and SRI International are part of two of the four teams. One will work on a project involving neuro-symbolic AI algorithm modeling, while another will aim to enhance the quantification of mission-specific tasks by creating an assurance framework and methodology.
Other organizations taking part in the projects are the University of California, Los Angeles, University of Central Florida, Monash University, Carnegie Mellon University, Systems & Technology Research and Johns Hopkins University Applied Physics Laboratory.
