Evaluating the roughness of structure–property relationships using pretrained molecular representations
David Graff, Edward O. Pyzer‐Knapp, Kirk E. Jordan, Eugene I. Shakhnovich, Connor W. Coley
Abstract
Pretrained molecular representations are often thought to provide smooth, navigable latent spaces; analysis by ROGI-XD suggests they are no smoother than fixed descriptor/fingerprint representations.
Topics & Concepts
Property (philosophy)Fingerprint (computing)Computer scienceArtificial intelligenceNatural language processingPattern recognition (psychology)PhilosophyEpistemologyComputational Drug Discovery MethodsMachine Learning in Materials ScienceProtein Structure and Dynamics