用智能代理自动设计高熵合金，准确率达94.66%。
From Phase Prediction to Phase Design: A ReAct Agent Framework for High-Entropy Alloy Discovery
- 构建基于推理与行动的LLM代理，自主迭代优化合金成分。
- 对三种晶相的重发现率最高达38%，比随机搜索快22.8倍。
- 兼顾文献密集区与未探索区域，适合新材料发现与工业研发。
发现能稳定形成目标晶体相的高熵合金（HEA）成分是一个高维逆向设计问题，传统试错法和仅正向预测的机器学习模型难以高效解决。本文提出一种基于ReAct（推理+行动）框架的大型语言模型代理，通过查询基于4,753条实验记录训练的校准XGBoost代理模型，实现成分的自主提议、验证与迭代优化，整体准确率达94.66%（宏平均F1为0.896）。相比贝叶斯优化与随机搜索基线，该代理在面心立方（FCC）、体心立方（BCC）及双相（BCC+FCC）相的描述符空间重发现率分别达到38%、18%和38%（曼-惠特尼检验p ≤ 0.039），其建议成分距离实验相流形仅2.4–22.8倍于随机搜索。消融实验表明，领域先验促使代理从已知合金回忆转向成分多样性探索——无先验代理更依赖文献密集族系，而完整提示代理能深入未被充分研究区域（以BCC+FCC为例，唯一性比率1.0对比0.39）。斯皮尔曼相关分析显示，代理推理与真实相分布显著一致（ρ = 0.736，p = 0.004，BCC相）。本工作确立了基于大模型的智能体推理在无梯度优化中的系统性、可解释性与相流形感知能力，为逆向合金设计提供新范式。
原文摘要 · Abstract (English)
Discovering high-entropy alloy (HEA) compositions that reliably form a target crystal phase is a high-dimensional inverse design problem that conventional trial-and-error experimentation and forward-only machine learning models cannot efficiently solve. Here we present a ReAct (Reasoning + Acting) LLM agent that autonomously proposes, validates, and iteratively refines HEA compositions by querying a calibrated XGBoost surrogate trained on 4,753 experimental records across four phases (FCC, BCC, BCC+FCC, BCC+IM), achieving 94.66\% accuracy (F1 macro = 0.896). Against Bayesian optimisation (BO) and random search baselines, the full-prompt agent achieves descriptor-space rediscovery rates of 38\%, 18\%, and 38\% for FCC, BCC, and BCC+FCC (Mann--Whitney $p \leq 0.039$), with proposals lying 2.4--22.8$\times$ closer to the experimental phase manifold than random search. An ablation reveals that domain priors shift the agent from landmark-alloy recall toward compositionally diverse exploration -- an uninformed agent scores higher rediscovery by concentrating on literature-dense families, while the full-prompt agent explores underrepresented space (unique ratio 1.0 vs.\ 0.39 for BCC+FCC). These regimes represent distinct criteria: proximity to known literature versus genuine discovery. Spearman analysis confirms agent reasoning is statistically aligned with empirical phase distributions ($ρ= 0.736$, $p = 0.004$ for BCC). This work establishes LLM-guided agentic reasoning as a principled, transparent, and manifold-aware complement to gradient-free optimisation for inverse alloy design.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。