Researchers at Rutgers University have discovered that learning can be mimicked in synthetic matter.
Learning is a fundamental feature of intelligence, and the new finding could have major implications for algorithm development in artificial intelligence (AI).
The new study was published in the journal PNAS .
Basic Characteristics of Man
The fundamental feature of learning in humans has inspired the development of many artificial intelligence technologies and enables them to adapt to changing conditions and environments.
However, AI often focuses on imitating human logic. The researchers’ new discovery offers a way to mimic human cognition in devices that can learn, remember, and make decisions in ways similar to humans and our brains.
By creating this in solid form, new algorithms in artificial intelligence and neuromorphic computing can be developed with the flexibility to address uncertainties, contradictions, and other similar aspects present in our daily lives.
Neuromorphic computing creates artificial nervous systems to transmit electrical signals that mimic brain signals and does so to mimic the general neural structure and functioning of the human brain.
Researchers at Rutgers were joined by colleagues at Purdue, the University of Georgia and Argonne National Laboratory.
The Role of Nickel Oxide
Together, the researchers examined how the electrical conductivity of a special type of insulating material, nickel oxide, responded after its environment was changed many times over various periods.
Subhashish Mandal is a postdoctoral associate in the Department of Physics and Astronomy at Rutgers-New Brunswick.
“The goal was to find a material whose electrical conductivity could be tuned by modulating the concentration of atomic defects with external stimuli such as oxygen, ozone and light,” Mandal said. “We studied how this material behaves when we mix the system with oxygen or hydrogen and, most importantly, how external stimulation changes the electronic properties of the material.”
One of the researchers’ findings was that the material was unable to respond fully when gas stimuli changed rapidly.
Instead, he remained in an uncertain state in both environments as his response began to wane.
The researchers then introduced a noxium stimulus, such as ozone, and the material responded more strongly before decreasing again.
“The most interesting aspect of our results is that they show universal learning features such as habit and sensitivity that we generally find in living species,” Mandal said. Said.
“These material properties could inspire new algorithms for artificial intelligence. Just as the collective motion of birds or fish has inspired artificial intelligence, we believe the collective behavior of electrons in a quantum solid could do the same in the future.”
“The growing field of artificial intelligence requires hardware that can accommodate adaptive memory capabilities beyond those used in today’s computers,” he continued. “We found that nickel oxide insulators, which have historically been limited to academic pursuits, may be interesting candidates to test in the future for brain-inspired computers and robotics.”
