AI model improves prediction of fusion energy pulse timing and shape
Decomposition-Guided Diffusion Language Models for Inertial Confinement Fusion Prediction
Machine Learning
Summary
Inertial confinement fusion is a method to create clean energy but experiments are very expensive, so accurate computer models are important. The authors developed a new AI model called ICF-DLM that predicts the detailed neutron pulse created by fusion experiments from input laser and target settings. Their model breaks down the problem into parts like total energy yield and peak timing, which helps it handle tricky timing and low data. It performs better than previous AI methods on both simulated and real fusion data. This approach might also help other science problems with limited data and rare events.
Inertial confinement fusionAI surrogate modelsTime-series predictionNeutron-rate waveformPeak timingLow data regimeDiffusion language modelAutoregressive modelsReward optimizationNational Ignition Facility
Authors
Xiang Zhang, Varchas Gopalaswamy, Rahman Ejaz, Riccardo Betti, Dongfang Liu
Abstract
Inertial confinement fusion (ICF) is a leading pathway toward clean energy, but each shot at the National Ignition Facility costs on the order of one million dollars, making accurate AI surrogates a high-value target. We study exogenous-driven ICF waveform prediction, where a 512-step neutron-rate diagnostic must be inferred directly from a laser pulse and target design parameters, with no historical response observed. The regime stresses standard time-series predictors with temporal sparsity (picosecond peak in a nanosecond window), input-output scale mismatch (under 300 real shots), and peak sensitivity (picosecond timing). We propose ICF-DLM, to our knowledge the first LM-based ICF predictor, combining (i) a physics-typed decomposition into yield $Y_{DT}$, peak timing $t_{\mathrm{peak}}$, and local waveform $w_{\mathrm{local}}$; (ii) bidirectional denoising that defers commitment to peak location; and (iii) a physics-driven PPO reward re-injecting metric structure across numeric tokens. On ICFBench (50K simulations + 232 experimental shots), ICF-DLM cuts peak-timing error from 11.6 to 9.2 steps over a matched autoregressive LLaMA-3-8B and outperforms classical sequence models and LLM-based time-series predictors. Beyond ICF, the recipe shows potential to address science domains with low data and sparse events.