Thought AI works year-round without a break? Well, think again!
Watkins and her team, at the Los Alamos National Laboratory, found that continuous working AI network simulations become unstable after being exposed to continuous learning processes. But when they are exposed to waves that a human brain experiences during sleep, they become stable.
"It was as though we were giving the neural networks the equivalent of a good night's rest," said Watkins. This breakthrough came when Watkins’ team tried to develop a neural network similar to humans and other beings.
“We study spiking neural networks, which are systems that learn much as living brains do," said Los Alamos National Laboratory computer scientist Yijing Watkins. "We were fascinated by the prospect of training a neuromorphic processor in a manner analogous to how humans and other biological systems learn from their environment during childhood development."
"The issue of how to keep learning systems from becoming unstable really only arises when attempting to utilize biologically realistic, spiking neuromorphic processors or when trying to understand biology itself," said Los Alamos computer scientist and study coauthor Garrett Kenyon. "The vast majority of machine learning, deep learning, and AI researchers never encounter this issue because in the very artificial systems they study they have the luxury of performing global mathematical operations that have the effect of regulating the overall dynamical gain of the system.
The researchers are planning to use the same principle on Loihi neuromorphic chips made by Intel to confirm whether AI brains need to sleep or not. If it proves to be true, then we can expect other intelligent machines to come in the future.