[ad_1]
Artificial intelligence (AI) can help develop a safe, clean and virtually unlimited fusion energy for power generation, according to scientists.
A team, made up of researchers from Princeton University and Harvard University, applies extensive learning to predict sudden disturbances that can stop fusion reactions and damage donut-shaped tokamaks or devices. who host these reactions.
In-depth learning is a powerful new version of the machine-learning form of machine learning, according to findings published in Nature magazine.
"This research opens a promising new chapter in efforts to bring unlimited energy to Earth," said Steven Cowley, director of the Princeton Plasma Physics Laboratory (PPPL) of the US Department of Energy. (DOE).
"Artificial intelligence is exploding in all sciences and is now starting to contribute to the global quest for fusion energy," Cowley said in a statement.
Fusion, which drives the sun and the stars, is the fusion of light elements in the form of plasma – the hot and charged state of matter, composed of free electrons and atomic nuclei – which generates energy.
Scientists seek to replicate the fusion on Earth to obtain an abundant source of energy for the production of electricity.
"Artificial intelligence is the most intriguing field of scientific growth at the moment, and associating it with the science of fusion is very exciting," said William Tang, the author. one of the leading research physicists of the PPPL.
"We have accelerated the ability to predict with great precision the most dangerous challenge of clean fusion energy," Tang said.
Unlike traditional software, which perform the prescribed instructions, deep learning learns from mistakes.
Neural networks, interconnected node layers – mathematical algorithms – that are "parameterized" or weighted by the program to give the desired shape to reality are what appear to be magical.
For any given input, the nodes seek to produce a specified output, such as a correct identification of a face or accurate forecasts of a disturbance.
The training starts when a node fails to perform this task: the weights automatically adjust for new data until the correct output is obtained.
[ad_2]
Source link