A team of researchers at Radboud University in the Netherlands claims to have created an intelligent material that could help develop a functional “quantum brain”, mimicking the behavior of neurons and the synapses between them.
Scientists have, in fact, been looking for a long time for a type of material that can be learned by physically changing itself, in the same way that our brains do. Hardware can be physically reconfigured as it processes and stores new information, similar to how the human brain works. For artificial intelligence systems to work efficiently, the computer running the system must be able to learn new patterns from the environment and recognize previously learned ones .
We can currently see this functionality using machine learning software, which stores and processes data on separate hard drives. However, because these systems consume so much power, scientists recognize that they need to develop new hardware that can produce the same results with less power. But this proposed quantum brain has the ability to store and manage data just like artificial intelligence but with a unique and innovative material. They have shown that they can model and interconnect a network of individual atoms.
First steps
They found that by building a network of cobalt atoms placed on black phosphorus they were able to build a material that stores and processes information in a similar way to the brain and, even more surprising, adapts itself. By creating custom sets of these atoms, the scientists found that the firing behavior of these sets mimics the behavior of a brain-like model used in artificial intelligence. In doing so , they created the smallest synapse known to date.
According to experts, who publish their work in the journal Nature Nanotechnology , it could help create a new generation of computers and reduce carbon dioxide emissions in the digital economy.
In 2018, the same team of scientists demonstrated that it was possible to store bits of information in a cobalt atom and induce the ‘firing’ by applying voltage. After firing, the atom moves between two electronic states (between the values of 0 and 1) similar to a neuron in the brain.
“By stimulating the material for a longer period of time with a certain voltage, we were very surprised to see that the synapses actually changed,” said project leader Alexander Khajetoorians, professor of scanning probe microscopy at Radboud University. “The material adapted its reaction based on the external stimuli it received. It taught itself.”
The next target
The researchers plan to build a larger network of atoms, as well as dive into new “quantum” materials that can be used for these purposes. In addition, they must understand why the network of atoms behaves as it does. “We are in a state where we can begin to relate fundamental physics with concepts in biology, such as memory and learning,” the authors state.
Referencia: An atomic Boltzmann machine capable of self-adaption, Nature Nanotechnology (2021). DOI: 10.1038/s41565-020-00838-4 , www.nature.com/articles/s41565-020-00838-4
