Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study systematically evaluates the robustness of plasma diagnostic machine learning models using the TokaMark dataset, which includes 11,573 shots from the MAST tokamak. The research highlights that while sequence models like LSTM suffer dramatically from disruption-proximate sensor failures, statistical models such as XGBoost maintain stability under similar conditions. Key findings reveal that forward-fill imputation can significantly mitigate random dropout effects, but fails to address degradation caused by corruption at the end of the observation window, emphasizing the critical role of plasma current diagnostics across architectures.
Disruption-proximate sensor failures can cause LSTM models to collapse, while XGBoost remains surprisingly stable, challenging assumptions about model robustness in fusion diagnostics.
Plasma diagnostic models for tokamak fusion devices are almost universally evaluated on clean, complete sensor data. In practice, fusion diagnostics fail regularly: acquisition systems start late, individual sensors die, and signal dropouts cluster precisely when a plasma disruption is approaching. We present the first systematic robustness benchmark for plasma diagnostic ML using the TokaMark dataset of 11,573 MAST shots, evaluating XGBoost, LSTM, Transformer, and the TokaMark CNN baseline across six physically-grounded failure scenarios and three imputation strategies. We introduce the Robustness Score (RS) for standardized cross-architecture comparison. Our central finding is that disruption-proximate sensor failure (corruption injected in the final window timesteps) collapses sequence model performance (LSTM +212% NRMSE) while a statistical feature model remains comparatively stable (XGBoost +37%). Forward-fill imputation eliminates nearly all degradation from random dropout for sequence models (LSTM +57% to ~0%), but offers little help when the end of the window is corrupted. Shot-level alarm evaluation using ground-truth disruption timestamps reveals that LSTM alarm detection collapses to TPR=0.00 under proximate sensor failure, while mean-fill imputation recovers it to TPR=1.00, a reversal of the pattern observed in NRMSE. Plasma current emerges as the single most critical diagnostic across all architectures (+73% to +140% upon removal). Code, data, and trained checkpoints are available at https://github.com/Neerav-Gupta/tokamark-robustness.