Artificial Brains: Machines are becoming more and more like us humans – especially when it comes to regeneration. A recent study found that. Accordingly, artificial brains also need beneficial and stimulating sleep to remain productive. Artificial intelligence – AI for short – is used in a wide variety of areas of our social life.
Whether in medical diagnostics, traffic monitoring, or the manufacturing industry: specific components, also known as neurons, constantly receive a large amount of data based on which the specific problem is to be solved.
A neural network analyzes the interactions between these neurons and determines various behavioral patterns, which it constantly monitors. If these patterns help solve a problem, they are adopted as human brains work. If the amount of data supplied increases, the number of neurons also increases – just like in a biological neuron. But not only this finding shows that an AI resembles us, humans.
Artificial Brains: With Wave-Like Sleep Phases For Recovery
A US science team from Los Alamos National Laboratory is convinced that artificial intelligence almost certainly needs a phase of rest. The basis for this finding is an experiment in which the team tried to teach neural networks the learning process of biological systems. In doing so, they confronted the networks with unsupervised dictionary training. The networks had to classify the objects without falling back on examples.
During the wide-ranging experiment, the scientists found that the network simulations worked unstably after an unsupervised learning phase without a break. When the researchers forced wave-like states on the networks, as our brains experience during sleep, the networks worked stably again and regenerated.
Artificial Brains Stimulated With Sounds
Previously, the researchers tried to support the networks with different noises, like those you know when tuning into a radio. Waves similar to Gaussian noise with wide amplitudes and frequencies represented the most significant positive effect.
The researchers confirmed their thesis that this noise comes close to what is known as slow-wave sleep, which biological neurons also go through. The test results led the research team to conclude that this slow-wave sleep gives the neural networks more stability while preventing hallucinations.
Will Intel Chips Retire In The Future?
In the next step, the researchers want to implement the algorithm they developed in Intel’s neuromorphic Loihi chip. Here, too, they rely on the “sleep” principle. This is how the chip should succeed in using and processing data collected by a silicon retina camera in real-time.
If these results are as successful as in the test, we can assume in the future that androids and other artificial bits of intelligence will now and then take a deep nap. A neural network analyzes the interactions between these neurons and determines various behavioural patterns, which it constantly monitors. If these patterns help solve a problem, they are adopted as human brains work.
Previously, the researchers tried to support the networks with different noises, like those you know when tuning into a radio. Waves similar to Gaussian noise with wide amplitudes and frequencies represented the most significant positive effect. If the amount of data supplied increases, the number of neurons also increases – just like in a biological neuron. But not only this finding shows that an AI resembles us, humans.
ALSO READ: 4 Reasons Why iPhone 12 Comes Without Earpods & Charging Connector
