The researchers at MIT claim to have created analog synapses that are one million times faster than the human brain.
Digital and analog chips need the same things. According to a press release, these resistors can be used to create a network of analog synapses.
Ultra-fast and efficient, these analog synapses aren't just that. Digital neural networks use more and more energy, increasing their carbon footprint considerably.
The researchers hope that their findings will help advance the field of deep learning.
The researchers were able to reach the speed of the human brain with the use of high tech glass instead of the usual organic medium.
Senior author and professor of nuclear science Ju Li said in the statement that the action potential in biological cells rises and falls with a time scale of milliseconds. "Here we apply up to ten volts across a special solid glass film that conducts protons without damaging it."
The stronger the field, the quicker the ionic devices.
It is possible for the protons to travel at ludicrous speeds while being incredibly energy- efficient due to the fact that it is resistant to high voltages.
It's easy to make and common, making it a practical option.
"After you have an analog processor, you will no longer be training networks everyone else is working on," said the lead author. You will be training networks that are so complex that no one else will be able to match them.
He said that this is not a faster car and that it is a spaceship.
Artificial intelligence is starting to be used by authors to quickly Churn out novels.
All Rights Reserved © 2024
