Global geometry of chemical graph neural network representations in terms of chemical moieties
Amer Marwan El-Samman, Incé A. Husain, Mai Huynh, Stefano De Castro, Brooke Morton, Stijn De Baerdemacker
Abstract
The embedding vectors from a Graph Neural Network trained on quantum chemical data allow for a global geometric space with a Euclidean distance metric. Moieties that are close in chemical sense, are also close in Euclidean sense.
Topics & Concepts
GraphArtificial neural networkGeometryComputer scienceArtificial intelligenceTheoretical computer scienceMathematicsComputational Drug Discovery MethodsMachine Learning in Materials ScienceMetabolomics and Mass Spectrometry Studies