Your browser doesn't support javascript.
loading
Composite Graph Neural Networks for Molecular Property Prediction.
Bongini, Pietro; Pancino, Niccolò; Bendjeddou, Asma; Scarselli, Franco; Maggini, Marco; Bianchini, Monica.
Afiliación
  • Bongini P; Department of Information Engineering and Mathematics, University of Siena, 53100 Siena, Italy.
  • Pancino N; Department of Information Engineering and Mathematics, University of Siena, 53100 Siena, Italy.
  • Bendjeddou A; Department of Information Engineering and Mathematics, University of Siena, 53100 Siena, Italy.
  • Scarselli F; Department of Information Engineering and Mathematics, University of Siena, 53100 Siena, Italy.
  • Maggini M; Department of Information Engineering and Mathematics, University of Siena, 53100 Siena, Italy.
  • Bianchini M; Department of Information Engineering and Mathematics, University of Siena, 53100 Siena, Italy.
Int J Mol Sci ; 25(12)2024 Jun 14.
Article en En | MEDLINE | ID: mdl-38928289
ABSTRACT
Graph Neural Networks have proven to be very valuable models for the solution of a wide variety of problems on molecular graphs, as well as in many other research fields involving graph-structured data. Molecules are heterogeneous graphs composed of atoms of different species. Composite graph neural networks process heterogeneous graphs with multiple-state-updating networks, each one dedicated to a particular node type. This approach allows for the extraction of information from s graph more efficiently than standard graph neural networks that distinguish node types through a one-hot encoded type of vector. We carried out extensive experimentation on eight molecular graph datasets and on a large number of both classification and regression tasks. The results we obtained clearly show that composite graph neural networks are far more efficient in this setting than standard graph neural networks.
Asunto(s)
Palabras clave
Texto completo
Añadir a Mi BVS
Imprimir
XML
PubMed Links
Buscar en Google

Texto completo: 1 Colección: 01-internacional Base de datos: MEDLINE Asunto principal: Redes Neurales de la Computación Idioma: En Revista: Int J Mol Sci Año: 2024 Tipo del documento: Article País de afiliación: Italia Pais de publicación: Suiza
Texto completo
Añadir a Mi BVS
Imprimir
XML
PubMed Links
Buscar en Google

Texto completo: 1 Colección: 01-internacional Base de datos: MEDLINE Asunto principal: Redes Neurales de la Computación Idioma: En Revista: Int J Mol Sci Año: 2024 Tipo del documento: Article País de afiliación: Italia Pais de publicación: Suiza