Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

Machine learning program to help bring the power of the sun to Earth

A new artificial intelligence (AI) program is getting closer to reproduce the process of fusion that powers the sun, according to new research from Princeton Plasma Physics Laboratory (PPPL).

The project, named “Accelerated Deep Learning Discovery in Fusion Energy Science”, is being developed by PPPL together with the US Department of Energy (DOE).

Fusion works by combining light elements in reactions that generate enormous amounts of energy. The researchers involved in the study aim to reproduce this process artificially to create a virtually inexhaustible supply of electric power.

Our research will utilise capabilities to accelerate progress that can only come from the deep learning form of artificial intelligence

William Tang

To control fusion processes on Earth scientists use tokamaks, a type of magnetic confinement devices. However, this is a volatile environment, and disruptions can halt the reactions and damage the tokamaks.

The breakthrough element of the research lies in the ability of the program to predict and tame such disruptions by using machine learning. Set to be completed in 2021, the system will be capable of performing a quintillion calculations per second, 50-to-100 times faster than the most powerful supercomputers today.

“Our research will utilise capabilities to accelerate progress that can only come from the deep learning form of artificial intelligence”, said William Tang, a principal research physicist at PPPL and professor in the Department of Astrophysical Sciences at Princeton University.

The program will be able to perform such advanced calculations because of the joint technologies that power it. It will utilise the database of the Joint European Torus (JET) in the United Kingdom, the largest and most powerful tokamak in the world today. It will also be able to access data from Tiger, a high-performance Princeton University cluster equipped with advanced image-resolution GPUs and great deep learning abilities, the Titan supercomputer at Oak Ridge National Laboratory and the Tsubame-III supercomputer in Tokyo, Japan.

SHARE THIS ARTICLE:
BY SHACK15

Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter