Scientists from DeepMind have found another real-world application for their artificial intelligence software.

London Artificial Intelligence Laboratory owned Alphabetannounced on Wednesday that it had taught the artificial intelligence system to monitor and sculpt superheated plasma inside a nuclear fusion reactor.

Nuclear fusion, the process that feeds the stars of the universe, involves the splitting and combining of hydrogen, which is a common element of seawater.

The process, which emits huge amounts of energy, is touted as a potentially limitless source of clean energy, but a number of technical challenges still need to be overcome.

Here on Earth, scientists are using tokamaks – donut-shaped vessels surrounded by electromagnetic coils – to try to recreate the fusion reactions that take place in space.

The magnets in these tacomas are used to “hold” volatile hydrogen plasma, which is hotter than the sun’s core. Controlling magnetic coils now requires several layers of sophisticated control systems.

Martin Riedmiller, head of the control group at DeepMind and co-author of Fr. article published in the journal Nature on Wednesday told CNBC it was a “really difficult” control issue.

Together with the Swiss Plasma Center at EPFL, University of Lausanne, Switzerland, DeepMind said it has developed an artificial intelligence system with reinforcement that can control magnets and change their voltage thousands of times per second.

Reinforcing learning, the AI ​​training method to which DeepMind is particularly tuned, involves programming the AI ​​to perform certain actions to maximize its chances of earning a reward in a particular situation. In other words, the algorithm “learns” to perform a task by looking for these pre-programmed rewards.

The unnamed AI DeepMind, developed on a virtual simulator, has been used about 100 times on a takomak at the Swiss Plasma Center, known as the Variable Configuration Tokamak. He operated the magnets in the tokamak for two seconds, which is the maximum amount of time the reactor can run before it overheats.

About 10-20 people with DeepMind worked on the AI ​​system along with about 5-10 people with EPFL.

“In my opinion, synthesis is one of the most fundamental sources of energy we have in the world,” said Federico Felici, a researcher at the Swiss Plasma Center, to CNBC. “Once we really master this technology … it’s a huge achievement because you will be able to have almost limitless energy in predictable times.”

Damien Ernst, a professor at the University of Liege who was not involved, called the study one of the most important examples of reinforced learning to date, adding that it “could significantly accelerate the development of fusion reactors and, ultimately, our ability to fight global changes. ”

DeepMind has decided to hack into artificial general intelligence, often referred to as the Holy Grail II.

The company began by developing AI systems that could master games such as chess and th. Now he wants to apply more of his technology to real-world applications and science.

Although Google has found the use of AI DeepMind, its technology has not found widespread use elsewhere.

DeepMind CEO Demis Khasabis said in a statement that the company had demonstrated the potential of AI to accelerate scientific progress and open up new areas of research in biology, chemistry, mathematics and now physics.

DeepMind employs about 1,000 people worldwide, including some of the world’s leading AI researchers who can earn more than $ 1 million a year. These top people, who often have doctorates in countries like Oxford, Cambridge, Stanford and MIT, can make that kind of money because they’re also looking for big technology companies like Facebook, Apple, Amazon and Microsoft.

Scientists from DeepMind have found another real-world application for their artificial intelligence software.

London Artificial Intelligence Laboratory owned Alphabetannounced on Wednesday that it had taught the artificial intelligence system to monitor and sculpt superheated plasma inside a nuclear fusion reactor.

Nuclear fusion, the process that feeds the stars of the universe, involves the splitting and combining of hydrogen, which is a common element of seawater.

The process, which emits huge amounts of energy, is touted as a potentially limitless source of clean energy, but a number of technical challenges still need to be overcome.

Here on Earth, scientists are using tokamaks – donut-shaped vessels surrounded by electromagnetic coils – to try to recreate the fusion reactions that take place in space.

The magnets in these tacomas are used to “hold” volatile hydrogen plasma, which is hotter than the sun’s core. Controlling magnetic coils now requires several layers of sophisticated control systems.

Martin Riedmiller, head of the control group at DeepMind and co-author of Fr. article published in the journal Nature on Wednesday told CNBC it was a “really difficult” control issue.

Together with the Swiss Plasma Center at EPFL, University of Lausanne, Switzerland, DeepMind said it has developed an artificial intelligence system with reinforcement that can control magnets and change their voltage thousands of times per second.

Reinforcing learning, the AI ​​training method to which DeepMind is particularly tuned, involves programming the AI ​​to perform certain actions to maximize its chances of earning a reward in a particular situation. In other words, the algorithm “learns” to perform a task by looking for these pre-programmed rewards.

The unnamed AI DeepMind, developed on a virtual simulator, has been used about 100 times on a takomak at the Swiss Plasma Center, known as the Variable Configuration Tokamak. He operated the magnets in the tokamak for two seconds, which is the maximum amount of time the reactor can run before it overheats.

About 10-20 people with DeepMind worked on the AI ​​system along with about 5-10 people with EPFL.

“In my opinion, synthesis is one of the most fundamental sources of energy we have in the world,” said Federico Felici, a researcher at the Swiss Plasma Center, to CNBC. “Once we really master this technology … it’s a huge achievement because you will be able to have almost limitless energy in predictable times.”

Damien Ernst, a professor at the University of Liege who was not involved, called the study one of the most important examples of reinforced learning to date, adding that it “could significantly accelerate the development of fusion reactors and, ultimately, our ability to fight global changes. ”

DeepMind has decided to hack into artificial general intelligence, often referred to as the Holy Grail II.

The company began by developing AI systems that could master games such as chess and th. Now he wants to apply more of his technology to real-world applications and science.

Although Google has found the use of AI DeepMind, its technology has not found widespread use elsewhere.

DeepMind CEO Demis Khasabis said in a statement that the company had demonstrated the potential of AI to accelerate scientific progress and open up new areas of research in biology, chemistry, mathematics and now physics.

DeepMind employs about 1,000 people worldwide, including some of the world’s leading AI researchers who can earn more than $ 1 million a year. These top people, who often have doctorates in countries like Oxford, Cambridge, Stanford and MIT, can make that kind of money because they’re also looking for big technology companies like Facebook, Apple, Amazon and Microsoft.