Nanowire networks mimic the network structure of the human brain. But can they learn and remember like the human brain? New research shows they can.
The brain’s powerful information processing capabilities are largely due to the network of connections formed by neurons and synapses. While we understand much of how the brain works, some aspects such as higher cognitive functions such as learning and memory remain elusive.
A type of nanotechnology, nanowire networks (NWNs) are typically made from highly conductive silver wires that are invisible to the naked eye, covered with a plastic material and formed into a mesh. Nanowires self-assemble to form dynamic, complex networks that integrate memory and processing, similar to those found in the brain.
Now, an international team led by researchers from the University of Sydney has proven how similar the NWN is to the human brain.
“This network of nanowires is like a synthetic neural network, because the nanowires function like neurons and the places where they connect to each other are analogous to synapses,” said co-author of the study. Zdenka Kuncic said.
To examine how well the NWN is indicative of cognitive function, researchers controlled a version of the test used to assess working memory in humans. n– Backtest.
Humans are n– Backtesting may display a series of letters or a series of images in succession. For each item in the sequence, we need to determine if it matches the presented item.n‘ previous item.Ann n-back A score of 7 on average indicates that an item that appeared 7 items ago can be recognized.
For NWN, the researchers changed the n-back tests into implementable subtasks. To conduct the test, researchers routed her NWN to a desired location.
“What we’ve done here is to manipulate the voltages on the end electrodes to force them to reroute instead of letting the network do its own thing,” says Alon, lead author of the study. says Loeffler. “We forced the pathway to go where we wanted it to go.”
Researchers have found that directing the NWN’s path improves its storage capacity and accuracy.
“Once we implemented it, its memory had much higher accuracy and didn’t really diminish over time. It suggests we’ve found a way,” Loeffler said.
The proof was in the tests.when they dosed changed nBacktesting to the -NWN was able to “remember” the desired endpoint of the electrical circuit 7 steps back, comparable to human memory.
After continuously reinforcing the NWN, researchers say they reach a point where the memory becomes fixed and no further reinforcement is needed.
“It’s like the difference between long-term and short-term memory in our brain,” Kuncic said. “If you want to remember something for a long time, you have to keep training your brain to integrate it or it will fade away over time.”
The researchers say the study shows that NWNs function like the human brain and can be used to improve robotics and sensor devices that require rapid decision-making.
“In this study, we found that higher cognitive functions typically associated with the human brain can be emulated in non-biological hardware,” said Loeffler. “Our current research paves the way for replicating brain-like learning and memory in non-biological hardware systems, suggesting that the underlying properties of brain-like intelligence may be physical. It suggests that
The study was published in a journal scientific progress.
Source: University of Sydney
