Fig. 2From: Scalable training of graph convolutional neural networks for fast and accurate predictions of HOMO-LUMO gap in moleculesIllustration of a SMILES representation of a molecule (left), its corresponding molecular structure (center), and its corresponding molecular graph (right). The atoms that are not explicitly denominated in the molecular structure are carbon atoms and the hydrogen atoms that create a covalence bond with a carbon atom. All the hydrogen atoms are suppressed in the figure for the sake of brevity but they are treated as nodes of the molecular graph in GCNN modelsBack to article page