Deep reinforcement learning controls tokamak fusion plasma
Deep reinforcement learning controls tokamak fusion plasma is graded peer reviewed on whataifound.org, with the AI's role graded search scaffold.
A reinforcement-learning controller learned to shape and stabilise the magnetic confinement of real fusion plasma inside the TCV tokamak, holding configurations that are hard to sustain by hand.
- Verification
- Peer reviewed
- Autonomy
- Search scaffold
- Lab
- Google DeepMind
- Model
- DeepMind RL controller
- Field
- Physics
- Date
- 2022-02-16
- Notability
- 56 Wikipedia language editions
What was found
Working with EPFL's Swiss Plasma Center, DeepMind trained a deep reinforcement-learning agent to command the magnetic control coils of the TCV tokamak. Learning in a simulator against specified target shapes, the controller then ran on the real machine, holding elongated, snowflake and other configurations, including a 'droplet' state with two separate plasmas at once. It is one of the first times a single learned controller replaced the hand-engineered cascade normally required.
Novelty check
Magnetic plasma control is normally built from many separately engineered controllers. A single RL-learned controller running on a real tokamak was new; the Nature paper documents the comparison against conventional control.
Caveats
Demonstrated on the TCV research tokamak, not a power-producing reactor, and the policy was trained in simulation before transfer to hardware. This is a control-engineering advance, not a solution to fusion energy. Independent groups have since extended RL plasma control to other machines.
Sources
Community discussion
How this is graded
whataifound.org grades every entry on two axes: verification (how solid the result is, from a machine-checked proof down to refuted) and autonomy (how much the AI did versus its human collaborators). This finding is peer reviewed and search scaffold. Full definitions are in the methodology.
Cite this entry
whataifound.org (2022). Deep reinforcement learning controls tokamak fusion plasma. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2022-02-16-tokamak-plasma-control