Matches in SemOpenAlex for { <https://semopenalex.org/work/W3217657218> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W3217657218 abstract "Molecular property prediction is one of the fastest-growing applications of deep learning with critical real-world impacts. Including 3D molecular structure as input to learned models their performance for many molecular tasks. However, this information is infeasible to compute at the scale required by several real-world applications. We propose pre-training a model to reason about the geometry of molecules given only their 2D molecular graphs. Using methods from self-supervised learning, we maximize the mutual information between 3D summary vectors and the representations of a Graph Neural Network (GNN) such that they contain latent 3D information. During fine-tuning on molecules with unknown geometry, the GNN still generates implicit 3D information and can use it to improve downstream tasks. We show that 3D pre-training provides significant improvements for a wide range of properties, such as a 22% average MAE reduction on eight quantum mechanical properties. Moreover, the learned representations can be effectively transferred between datasets in different molecular spaces." @default.
- W3217657218 created "2021-12-06" @default.
- W3217657218 creator A5037753193 @default.
- W3217657218 creator A5056748708 @default.
- W3217657218 creator A5057871229 @default.
- W3217657218 creator A5064214996 @default.
- W3217657218 creator A5074504351 @default.
- W3217657218 creator A5088531553 @default.
- W3217657218 creator A5088533003 @default.
- W3217657218 date "2021-09-24" @default.
- W3217657218 modified "2023-09-26" @default.
- W3217657218 title "3D Infomax improves GNNs for Molecular Property Prediction" @default.
- W3217657218 hasPublicationYear "2021" @default.
- W3217657218 type Work @default.
- W3217657218 sameAs 3217657218 @default.
- W3217657218 citedByCount "0" @default.
- W3217657218 crossrefType "journal-article" @default.
- W3217657218 hasAuthorship W3217657218A5037753193 @default.
- W3217657218 hasAuthorship W3217657218A5056748708 @default.
- W3217657218 hasAuthorship W3217657218A5057871229 @default.
- W3217657218 hasAuthorship W3217657218A5064214996 @default.
- W3217657218 hasAuthorship W3217657218A5074504351 @default.
- W3217657218 hasAuthorship W3217657218A5088531553 @default.
- W3217657218 hasAuthorship W3217657218A5088533003 @default.
- W3217657218 hasConcept C111472728 @default.
- W3217657218 hasConcept C119857082 @default.
- W3217657218 hasConcept C120317606 @default.
- W3217657218 hasConcept C127162648 @default.
- W3217657218 hasConcept C127413603 @default.
- W3217657218 hasConcept C132525143 @default.
- W3217657218 hasConcept C138885662 @default.
- W3217657218 hasConcept C146978453 @default.
- W3217657218 hasConcept C152139883 @default.
- W3217657218 hasConcept C153402090 @default.
- W3217657218 hasConcept C154945302 @default.
- W3217657218 hasConcept C189950617 @default.
- W3217657218 hasConcept C204323151 @default.
- W3217657218 hasConcept C31258907 @default.
- W3217657218 hasConcept C41008148 @default.
- W3217657218 hasConcept C80444323 @default.
- W3217657218 hasConceptScore W3217657218C111472728 @default.
- W3217657218 hasConceptScore W3217657218C119857082 @default.
- W3217657218 hasConceptScore W3217657218C120317606 @default.
- W3217657218 hasConceptScore W3217657218C127162648 @default.
- W3217657218 hasConceptScore W3217657218C127413603 @default.
- W3217657218 hasConceptScore W3217657218C132525143 @default.
- W3217657218 hasConceptScore W3217657218C138885662 @default.
- W3217657218 hasConceptScore W3217657218C146978453 @default.
- W3217657218 hasConceptScore W3217657218C152139883 @default.
- W3217657218 hasConceptScore W3217657218C153402090 @default.
- W3217657218 hasConceptScore W3217657218C154945302 @default.
- W3217657218 hasConceptScore W3217657218C189950617 @default.
- W3217657218 hasConceptScore W3217657218C204323151 @default.
- W3217657218 hasConceptScore W3217657218C31258907 @default.
- W3217657218 hasConceptScore W3217657218C41008148 @default.
- W3217657218 hasConceptScore W3217657218C80444323 @default.
- W3217657218 hasLocation W32176572181 @default.
- W3217657218 hasOpenAccess W3217657218 @default.
- W3217657218 hasPrimaryLocation W32176572181 @default.
- W3217657218 isParatext "false" @default.
- W3217657218 isRetracted "false" @default.
- W3217657218 magId "3217657218" @default.
- W3217657218 workType "article" @default.