Toward Next-Generation Heterogeneous Catalysts: Empowering Surface Reactivity Prediction with Machine Learning

From Characterization to Discovery: Artificial Intelligence, Machine Learning and High-Throughput Experiments for Heterogeneous Catalyst Design

“…Applications focused on the prediction of structural and surface chemistry of catalysts are described below. Li and Peng (2024) generalize two complementary strategies regarding the descriptors or features used to train the ML algorithm: ML with ab initio features: based on fundamental theoretical calculations. The use of ab initio features in ML models can significantly enhance the prediction accuracy of molecular properties, such as binding energies, demonstrating that detailed quantum mechanical insights provide valuable information for machine learning predictions. ML with non- ab initio features: such as simplifications, heuristics, and elemental properties.…”

“…The applications described in this section contain elements of property prediction, as this is the main strategy for the screening of catalysts . However, the works cited focus on the discovery of new catalysts.…”