QuantumFusion Blog

DeepMind AI in Fusion

3/25/2022   QuantumFusion Blog Post #1   By: Akhil Surapaneni

My project to link Quantum and Nuclear Fusion is advancing at a fast rate in labs around the world, but a precursor to it could be a DeepMind AI controlling plasma inside a fusion reactor. Plasma inside a fusion reactor must be constantly monitored, and its magnetic fields constantly manipulated so that it doesn't touch the walls of a fusion reactor, damaging the machine and cooling the plasma. The DeepMind AI will be the one to constantly monitor and manipulate the plasma inside a fusion reactor. A team of scientists taught this AI to control and change the plasma and gave it command of the plasma in a Variable Configuration Tokamak (TCV). The AI was able to control the plasma in the reactor for 2 seconds, which was the whole length of time the TCV reactor could run.

This is eventually where my project, QuantumFusion, comes in. The relatively phenomenal success of the DeepMind AI proves we have much more space to grow and Quantum Computers can get us there. Using quantum machine learning we will be able to perfect an algorithm that a quantum computer can then use to control plasma. Quantum Computers also prove advantageous over DeepMind AI because of its ability to simulate data at a molecular level. Although DeepMind AI can control a reactor, it does not have the capability to understand plasma in the way Quantum Computers will. For these reasons, in my project, I want to use Quantum Computers as the mainframe to control and simulate the plasma in my Fusion Reactor.