Towards explainable message passing networks for predicting carbon dioxide adsorption in metal-organic frameworks
This work addresses the problem of predicting CO2 adsorption in MOFs for climate change mitigation, providing incremental insights into material design.
This paper developed a message passing neural network (MPNN) to predict CO2 adsorption in Metal-Organic Frameworks (MOFs). The model incorporates a soft attention mechanism to identify important substructures within MOFs contributing to the prediction.
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.