Human-in-the-Loop Meta Bayesian Optimization for Fusion Energy and Scientific Applications
For scientists in high-stakes, data-scarce fields like fusion energy, HL-MBO provides a practical framework that combines machine learning with human expertise to improve optimization efficiency.
HL-MBO integrates expert knowledge with meta-learned surrogate models and interpretable acquisition functions to accelerate optimization in data-scarce scientific domains, outperforming current BO methods on ICF energy yield optimization and benchmarks in molecular optimization and superconductivity.
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.