Inertial Confinement Fusion (ICF) holds transformative promise for sustainable, near-limitless clean energy, yet remains constrained by prohibitively high costs and limited experimental opportunities. This paper presents Human-in-the-Loop Meta Bayesian Optimization (HL-MBO), a framework that integrates expert knowledge with few-shot, uncertainty-aware machine learning to accelerate discovery in data-scarce, high-stakes scientific domains. HL-MBO introduces a meta-learned surrogate model with an expert-informed acquisition function to recommend candidate experiments. To foster trust and enable informed decisions, HL-MBO also provides interpretable explanations of its suggestions. We show HL-MBO outperforms current BO methods on ICF energy yield optimization, as well as benchmarks in molecular optimization and critical temperature maximization for superconducting materials.
@article{arxiv.2605.00068,
title = {Human-in-the-Loop Meta Bayesian Optimization for Fusion Energy and Scientific Applications},
author = {Ricardo Luna Gutierrez and Sahand Ghorbanpour and Ejaz Rahman and Varchas Gopalaswamy and Riccardo Betti and Vineet Gundecha and Aarne Lees and Soumyendu Sarkar},
journal= {arXiv preprint arXiv:2605.00068},
year = {2026}
}
Comments
Accepted at IJCAI 2026 (35th International Joint Conference on Artificial Intelligence)