Towards explainable message passing networks for predicting carbon dioxide adsorption in metal-organic frameworks

Abstract

Metal-organic framework (MOFs) are nanoporous materials that could be used to capture carbon dioxide from the exhaust gas of fossil fuel power plants to mitigate climate change. In this work, we design and train a message passing neural network (MPNN) to predict simulated CO 2 adsorption in MOFs. Towards providing insights into what substructures of the MOFs are important for the prediction, we introduce a soft attention mechanism into the readout function that quantifies the contributions of the node representations towards the graph representations. We investigate different mechanisms for sparse attention to ensure only the most relevant substructures are identified.

Accurate, interpretable predictions of materials properties within transformer language models

“… 80 , 81 , 82 The field is in its infancy, and therefore explainability-aware studies exploring graph neural networks in materials science are sparse and few. 83 , 84 , 85 …”