Improving Few- and Zero-Shot Reaction Template Prediction Using Modern Hopfield Networks
Philipp Seidl, Philipp Renz, Natalia Dyubankova, Paulo Neves, Jonas Verhoeven, Jörg K. Wegner, Marwin Segler, Sepp Hochreiter, Günter Klambauer
Abstract
≥ 3 in the retrosynthesis benchmark USPTO-50k. Code to reproduce the results is available at github.com/ml-jku/mhn-react.
Topics & Concepts
TemplateRetrosynthetic analysisComputer scienceGeneralizationBenchmark (surveying)Relevance (law)InferenceEncoding (memory)Representation (politics)Artificial intelligenceAlgorithmDependency (UML)Theoretical computer scienceMachine learningProgramming languageMathematicsPolitical sciencePoliticsGeographyMathematical analysisOrganic chemistryChemistryLawGeodesyTotal synthesisComputational Drug Discovery MethodsMachine Learning in Materials ScienceProtein Structure and Dynamics