Reinforcement learning agent trained in DIII-D tokamak simulator achieves 2.01 cm mean shape error on held-out data, tracks dynamic targets, and remains functional under 30% random sensor dropout with direct transfer to experimental shots.
Machine learning-based real-time kinetic profile reconstruction in DIII-D , volume =
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
ECHO deploys a real-time neural-network surrogate of TORBEAM combined with genetic optimization to control ECH deposition profiles on DIII-D, validated against ECE data despite gyrotron failures.
citing papers explorer
-
Dynamic Plasma Shape Control with Arbitrary Sensor Subsets
Reinforcement learning agent trained in DIII-D tokamak simulator achieves 2.01 cm mean shape error on held-out data, tracks dynamic targets, and remains functional under 30% random sensor dropout with direct transfer to experimental shots.
-
Robust Control of ECH Deposition Profiles on DIII-D
ECHO deploys a real-time neural-network surrogate of TORBEAM combined with genetic optimization to control ECH deposition profiles on DIII-D, validated against ECE data despite gyrotron failures.