AI in Nuclear Fusion Plasma Control Quiz

Answers and explanations appear as you take the quiz.

1. What do magnetic fields control in the tokamak confinement context described here?
   • Only the color of a decorative image
   • Only the retail electricity price
   • A guarantee that every device has the same temperature
   • The confinement, position and shape of charged plasma
2. What machine-learning method did DeepMind use to control the TCV tokamak?
   • Reinforcement learning
   • Supervised image classification
   • Simple linear regression
   • A rule-based expert system
3. Why does the cited TCV control work use a simulated environment for policy training?
   • It proves simulation and every physical device are identical
   • It provides trial-and-error experience before the evaluated hardware tests
   • It eliminates the need to validate transfer
   • It means no tokamak hardware was used
4. What is a plasma 'disruption' that AI tries to forecast?
   • A scheduling conflict
   • A power bill
   • A sudden collapse of the plasma that can damage the reactor
   • A software bug
5. What did the DeepMind controller output to steer the plasma?
   • Voltage commands for the magnetic coils
   • Tweets
   • Temperature readings only
   • Images of the plasma
6. On which experimental tokamak was the cited 2022 learned magnetic-control demonstration performed?
   • A proven commercial fusion power station
   • Only an unspecified future ITER power plant
   • A standard battery factory
   • TCV at the Swiss Plasma Center
7. At a uniform cadence of 10,000 control interactions per second, what interval separates interactions?
   • 10 milliseconds
   • 100 milliseconds
   • 0.1 milliseconds
   • 1 second
8. What does the reported TCV droplet configuration contain?
   • A completed commercial electric generator
   • Two plasmas inside the vessel simultaneously
   • Only a liquid-water cooling droplet with no plasma
   • An unlimited number of separate reactors