Open Catalyst

Open Catalyst is a research initiative focused on accelerating catalyst discovery using machine learning, specifically graph neural networks (GNNs). Current research emphasizes developing efficient GNN architectures, such as lightweight models and few-shot learning approaches, to overcome computational limitations and data scarcity in predicting catalyst-adsorbate interactions. This work leverages large datasets of simulated catalytic reactions to improve the accuracy and speed of predicting adsorption energies and reaction pathways, ultimately aiming to significantly reduce the time and cost associated with developing new catalysts for energy applications and beyond.