Predicting the enthalpy of formation of energetic molecules via conventional machine learning and GNN

Different ML models are used to map the enthalpy of formation from the molecular structure, and the impact of different feature representation methods and models on the results is explored. Among them, GNN models achieve impressive results.