Decoding Chemical Predictions: Group Contribution Methods for XAI

Graph Neural Networks (GNNs) have recently shown great promise for modeling chemical systems. However, beyond the accuracy and performance of these models, understanding their underlying mechanisms is also crucial. While many general GNN explainers exist, incorporating domain-specific knowledge can enhance the development of explainers tailored to chemical applications. In this study, we developed an explainability approach based on the well-established concept of group contributions. Our approach provides additional explanations without compromising model accuracy. Furthermore, our results indicate that different GNN models may learn distinct patterns from the molecules.