Researchers at USC and Beyond study AI’s ability to emulate lifelong learning.

Can a computer learn in the same way as people or animals?

Researchers in the field Artificial Intelligence are constantly trying to find out if a computer can ever be mistaken for a human. Since Alan Turing created the Turing Test, which determines whether a machine can exhibit intelligence that is equal to or indistinguishable with a human, AI has made great advances. Machines still have a hard time with a fundamental skill that comes naturally to humans and other living organisms: lifelong education. It’s about learning and adapting as we go along, without forgetting the previous task, or transferring intuitively knowledge from one area to another.

With the help of DARPA’s Lifelong Learning Machines program, USC Viterbi scientists have worked with colleagues from institutions around the U.S. to create a new resource that will be used for AI learning in the future. This resource defines how artificial systems are able to think, act, and adapt successfully in the real-world, just like living creatures.

The paper was co-authored by Alice Parker, Dean’s Professor of Electrical and Computer Engineering and Francisco Valero-Cuevas, Professor of Biomedical Engineering and of Biokinesiology and Physical Therapy at the University of Texas San Antonio. They also collaborated with Professor Dhireesha Kudithipudi and 22 other universities.