قالب وردپرس درنا توس
Home / Science / AI can help develop clean, boundless fusion energy artificially Intelligence | nuclear | Merger | AI | deep learning | neural network | Future | limitless

AI can help develop clean, boundless fusion energy artificially Intelligence | nuclear | Merger | AI | deep learning | neural network | Future | limitless



Artificial intelligence (AI) can help develop safe, clean and virtually unlimited fusion energy for generating electricity, scientists say.

A team that includes researchers from Princeton University and Harvard University conducts deep learning to predict sudden disturbances, stop fusion reactions, and damage the donut-shaped tokamaks or apparatuses that harbor the reactions.

Deep Learning is a powerful new version of AI's machine learning, as published in the journal Nature.

"This research opens a promising new chapter in the quest to bring unlimited energy to Earth," said Steven Cowley, director of the US Department of Energy's Princeton Plasma Physics Laboratory (PPPL).

"Artificial intelligence is exploding in the sciences and now it is beginning to contribute to the worldwide search for merger power," Cowley said in a statement.

Fusion, which drives the sun and the stars, is the Fus light elements in the form of plasma ̵

1; the hot, charged state of matter, which consists of free electrons and atomic nuclei – generate energy.

Scientists are trying to replicate the fusion on Earth for an abundant supply of energy for the production of energy.

"Artificial intelligence is currently the most intriguing field of scientific growth, and it is very exciting to marry it to fusion science," said William Tang, a senior research physicist at the PPPL.

Ability to predict with high accuracy the most dangerous challenge for the purification of fusion energy, "said Tang.

Unlike traditional software that performs prescribed instructions, deep learning learns from its mistakes.

To achieve magic are neural networks, layers of interconnected nodes – mathematical algorithms – that are "parameterized" or weighted by the program to form the desired output.

For any given input, the nodes seek training

starts when a node does not accomplish this task: the weights automatically adjust for new data until the correct output is obtained.


Source link