A joint diffusion approach to multi-modal inference in inertial confinement fusion
Michael S. Jones, Justin Kunimune, Daniel Casey, Bogdan Kustowski, Eugene Kur, Kelli Humbird
arXiv:2601.21006·physics.plasm-ph·Published 2026-01-28·Updated 2026-02-04
A combination of physics-based simulation and experiments has been critical to achieving ignition in inertial confinement fusion (ICF). Simulation and experiment both produce a mixture of scalar and images outputs, however only a subset of simulated data are available experimentally. We introduce a generative framework, called JointDiff, which enables predictions of conditional simulation input and output distributions from partial, multi-modal observations. The model leverages joint diffusion to unify forward surrogate modeling, inverse inference, and output imputation into one architecture. We train our model on a large ensemble of three-dimensional Multi-Rocket Piston simulations and demonstrate high accuracy, statistical robustness, and transferability to experiments performed at the National Ignition Facility (NIF). This work establishes JointDiff as a flexible generative surrogate for multi-modal scientific tasks, with implications for understanding diagnostic constraints, aligning simulation to experiment, and accelerating ICF design.
TopicsFluid Dynamics & Plasma Physics
Tagssurrogate-modeling
arXiv categoriesphysics.plasm-ph
arXiv abstract pagePDF