arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.05341physics.chem-phcs.LG
收藏

利用上下文学习进行催化的贝叶斯优化

Bayesian Optimization of Catalysis With In-Context Learning

FutureHouse Inc.(FutureHouse公司);University of Rochester(罗切斯特大学);Department of Chemical Engineering University of Rochester(罗切斯特大学化学工程系)

Mayk Caldas Ramos, Shane S. Michtavy, Marc D. Porosoff, Andrew D. White

AI总结:

本文提出BO-ICL,利用冻结LLM的上下文学习实现自然语言空间中的不确定性回归与贝叶斯优化,并在催化实验中快速筛选近优催化剂。

摘要:

大型语言模型(LLM)能够通过上下文学习,在零样本或少样本情况下执行准确分类。我们将这一能力扩展到带不确定性估计的回归任务,使用冻结的LLM(例如GPT-3.5、Gemini),从而无需显式模型训练或特征工程即可在自然语言中实现贝叶斯优化(BO)。我们将其应用于材料发现:把实验催化剂合成与测试流程表示为自然语言提示。材料发现中的一个关键挑战是需要表征次优候选物,这会拖慢进展。尽管BO能有效探索大型设计空间,但高斯过程等标准代理模型假设平滑性和连续性,而这一假设在多相催化等高度非线性领域并不成立。我们的任务无关BO工作流直接在语言空间中运行,克服了该问题,可在不需要结构或电子描述符的情况下产生可解释且可执行的预测。在水溶性和甲烷氧化偶联(OCM)等基准上,BO-ICL达到或超过高斯过程的表现。在针对逆水煤气变换(RWGS)反应的实时实验中,BO-ICL在6次迭代内从3,700个候选物池中识别出接近最优的多金属催化剂。我们的方法重新定义了材料表征并加速发现,在催化、材料科学和AI领域具有广泛应用。代码:https://github.com/ur-whitelab/BO-ICL。

英文摘要:

Large language models (LLMs) can perform accurate classification with zero or few examples through in-context learning. We extend this capability to regression with uncertainty estimation using frozen LLMs (e.g., GPT-3.5, Gemini), enabling Bayesian optimization (BO) in natural language without explicit model training or feature engineering. We apply this to materials discovery by representing experimental catalyst synthesis and testing procedures as natural language prompts. A key challenge in materials discovery is the need to characterize suboptimal candidates, which slows progress. While BO is effective for navigating large design spaces, standard surrogate models like Gaussian processes assume smoothness and continuity, an assumption that fails in highly non-linear domains such as heterogeneous catalysis. Our task-agnostic BO workflow overcomes this by operating directly in language space, producing interpretable and actionable predictions without requiring structural or electronic descriptors. On benchmarks like aqueous solubility and oxidative coupling of methane (OCM), BO-ICL matches or outperforms Gaussian processes. In live experiments on the reverse water-gas shift (RWGS) reaction, BO-ICL identifies near-optimal multi-metallic catalysts within six iterations from a pool of 3,700 candidates. Our method redefines materials representation and accelerates discovery, with broad applications across catalysis, materials science, and AI. Code: https://github.com/ur-whitelab/BO-ICL.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑