What is Google’s DeepMind and how is it used in nuclear fusion – Technology News, Firstpost
FP StaffFebruary 18, 2022 17:08:26 IST
In the race to create and contain carbon-free fusion energy, Alphabet’s artificial intelligence lab DeepMind makes the latest contribution.
In collaboration with the Swiss plasma center of EPFL – a university in Lausanne, Switzerland – the DeepMind AI has applied its algorithms to regulate the plasma in the nuclear fusion reactor, which is hotter than the surface of the sun, and maintain the temperature long enough to take energy out. That has all been nearly impossible until now.
Before we get into the scientific jargon, let’s break it down for you:
What is nuclear fusion?
The best and easiest to understand example of nuclear fusion energy is the sun. The process of nuclear fusion generates heat in the sun.
Creating fusion energy in labs has proven difficult because it consumes far more energy than it produces, making it widely useless as an energy source.
Existing nuclear power plants work on nuclear fission reactions that create energy by splitting atoms, a nuclear fusion reactor works the exact opposite of that, it releases energy by combining atoms.
Recently, the Joint European Torus (JET), a fusion reactor in Oxfordshire, UK, produced 59 megajoules of energy, equivalent to 11 megawatts of power, in a five-second period.
Also read: British scientists make major breakthrough in fusion energy: everything you need to know
The scientists built a process that allowed for the self-heating of matter when it is in a plasma state, using nuclear fusion, which could be an important step towards using nuclear fusion.
According to the Independent, the scientists took the hydrogen isotopes deuterium – which is found in seawater – and tritium, which is made in a reactor. They used the hydrogen isotopes to create a burning plasma.
Basically, the researchers were able to compress and heat a plasma, which is then heated by the reactions themselves, allowing the energy to sustain itself.
Due to the enormous gravity in the core of the sun, nuclear fusion is possible at a temperature of about 10 million Celsius. Since it is not possible to create such pressures on Earth, the temperatures must be much higher – above 100 million Celsius.
Since no material can withstand such a temperature, fusion is achieved in a superheated gas, or plasma, held in a doughnut-shaped magnetic field.
The Problems of Harvesting Nuclear Fusion Energy
Even though scientists have succeeded in creating nuclear fusion energy, it still faces a technical challenge – heating the plasma and holding it together to extract energy from it.
The process of confining and controlling the plasma can consume more energy than what is produced from it, making it a counterproductive process.
To put it in perspective, the recent experiment in the JET lab produced enough energy to boil 60 kettles of water for five seconds. Still, it was considered a major breakthrough.
Researchers have tried to limit nuclear fusion reactions and push them into various shapes that can yield maximum output using powerful magnetic coils.
In doing so, however, they must prevent the plasma from hitting the walls, which would not only damage the walls, but also waste the heat. So slow down the process of nuclear fusion.
What is Deep Mind?
DeepMind, a division of Google’s parent company, Alphabet, is responsible for developing general-purpose artificial intelligence technology. The technology takes input and learns from it through experience.
DeepMind claims that its system is not pre-programmed: it learns from experience and uses only raw pixels as data input.
Simply put, it used to be used for independent learning and playing games. When given the task of beating the library of Atari games, it learned to understand the games and over time, the AI was able to play the games better and more efficiently than humans.
The AI made headlines in 2016 when its AlphaGo program defeated Lee Sedol, the world champion of game “Go”, in a five-game match.
After proving its power playing video and board games, the AI has been used in healthcare to identify eye disease and kidney injuries, create computer programs and now advance in nuclear fusion research.
DeepMind’s contribution to nuclear fusion research
According to a report from Business Insider, the artificial intelligence lab and its co-researchers trained an algorithm on the Swiss Center’s simulator to hypothesize for itself how best to control the magnetic coils through gain learning.
This is where algorithms are effectively “rewarded” for achieving strong results.
It developed an architecture to preserve the plasma, sculpting it into different shapes while preserving separate plasma.
The algorithm was then applied to the real tokamak – a donut-shaped vacuum chamber with metal coils and a huge magnet to create sun-like conditions on Earth.
The report said the AI managed to manipulate the magnetic field for more than two seconds, just like it did in the simulator.
“This work is an important step in our understanding of how we can design new tokamaks that incorporate AI, and in the future we expect to see more and more sophistication in the use of reinforcement learning in the field,” said Ambrogio Fasoli, director of the Swiss Plasma Center, as quoted by Business Insider.
With input from agencies