Does a machine learning model actually gain an understanding of the material space? We answer this question in the affirmative on the example of the OptiMate model, a graph attention network trained to predict the optical properties of semiconductors and insulators. By applying the UMAP dimensionality reduction technique to its latent embeddings, we demonstrate that the model captures a nuanced and interpretable representation of the materials space, reflecting chemical and physical principles, without any user-induced bias. This enables clustering of almost 10,000 materials based on optical properties and chemical similarities. Beyond this understanding, we demonstrate how the learned material space can be used to identify more sustainable alternatives to critical materials in energy-related technologies, such as photovoltaics. These findings demonstrate the dual utility of machine learning models in materials science: Accurately predicting material properties while providing insights into the underlying materials space. The approach demonstrates the broader potential of leveraging learned materials spaces for the discovery and design of materials for diverse applications, and is easily applicable to any state-of-the-art machine learning model.
Sustainable Materials Discovery in the Era of Artificial Intelligence
Department of Civil and Environmental Engineering, Indian Institute of Technology Delhi(印度理工学院德里分校土木与环境工程系);Department of Civil and Environmental Engineering, Imperial College London(帝国理工学院伦敦分校土木与环境工程系);Department of Metallurgical and Materials Engineering, Indian Institute of Technology Madras(印度理工学院马德拉斯分校冶金与材料工程系);Department of Computer Science and Engineering, Chalmers University of Technology & University of Gothenburg(查尔姆斯理工大学与哥德堡大学计算机科学与工程系);Yardi School of Artificial Intelligence, Indian Institute of Technology Delhi(印度理工学院德里分校亚里学校人工智能系)
Accelerating Combinatorial Electrocatalyst Discovery with Bayesian Optimization: A Case Study in the Quaternary System Ni-Pd-Pt-Ru for the Oxygen Evolution Reaction
Leveraging Data Mining, Active Learning, and Domain Adaptation in a Multi-Stage, Machine Learning-Driven Approach for the Efficient Discovery of Advanced Acidic Oxygen Evolution Electrocatalysts
Pritzker School of Molecular Engineering, University of Chicago(芝加哥大学普利兹克分子工程学院);Data Science Institute, University of Chicago(芝加哥大学数据科学研究所);Argonne National Laboratory(阿贡国家实验室);North China Electric Power University(华北电力大学);Nanjing University(南京大学);The Hong Kong University of Science and Technology(香港科技大学);Department of Computer Science, University of Chicago(芝加哥大学计算机科学系)
Computational Design of Metal-Free Porphyrin Dyes for Sustainable Dye-Sensitized Solar Cells Informing Energy Informatics and Decision Support
Dinajpur Government College(迪纳杰布尔政府学院);UiT The Arctic University of Norway(挪威北极大学);Hajee Mohammad Danesh Science and Technology University(哈吉·穆罕默德·达内什科学与技术大学);Daffodil International University(达芬奇国际大学)
Revealing Hydroxide Ion Transport Mechanisms in Commercial Anion-Exchange Membranes at Nano-Scale from Machine-learned Interatomic Potential Simulations
Technische Universität Ilmenau(伊尔梅瑙工业大学);Fraunhofer-Institut für Keramische Technologien und Systeme IKTS(弗劳恩霍夫陶瓷技术与系统研究所)