Human-in-the-Loop Meta Bayesian Optimization for Fusion Energy and Scientific Applications

用于聚变能源与科学应用的人机协同元贝叶斯优化

Inertial Confinement Fusion (ICF) holds transformative promise for sustainable, near-limitless clean energy, yet remains constrained by prohibitively high costs and limited experimental opportunities. 惯性约束聚变（ICF）在实现可持续、近乎无限的清洁能源方面具有变革性的前景，但目前仍受限于高昂的成本和有限的实验机会。

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. 本文提出了一种人机协同元贝叶斯优化（HL-MBO）框架，该框架将专家知识与少样本、具备不确定性感知能力的机器学习相结合，旨在加速数据稀缺、高风险科学领域的研究发现。

HL-MBO introduces a meta-learned surrogate model with an expert-informed acquisition function to recommend candidate experiments. HL-MBO 引入了一种元学习代理模型，并结合了专家引导的采集函数，用于推荐候选实验方案。

To foster trust and enable informed decisions, HL-MBO also provides interpretable explanations of its suggestions. 为了建立信任并支持明智的决策，HL-MBO 还为其建议提供了可解释的说明。

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. 研究表明，在 ICF 能量产出优化，以及分子优化和超导材料临界温度最大化等基准测试中，HL-MBO 的表现均优于现有的贝叶斯优化（BO）方法。