Cargando…

Federated learning of molecular properties with graph neural networks in a heterogeneous setting

Chemistry research has both high material and computational costs to conduct experiments. Intuitions are interested in differing classes of molecules, creating heterogeneous data that cannot be easily joined by conventional methods. This work introduces federated heterogeneous molecular learning. Fe...

Descripción completa

Detalles Bibliográficos
Autores principales: Zhu, Wei, Luo, Jiebo, White, Andrew D.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Elsevier 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9214329/
https://www.ncbi.nlm.nih.gov/pubmed/35755872
http://dx.doi.org/10.1016/j.patter.2022.100521
_version_ 1784730990840643584
author Zhu, Wei
Luo, Jiebo
White, Andrew D.
author_facet Zhu, Wei
Luo, Jiebo
White, Andrew D.
author_sort Zhu, Wei
collection PubMed
description Chemistry research has both high material and computational costs to conduct experiments. Intuitions are interested in differing classes of molecules, creating heterogeneous data that cannot be easily joined by conventional methods. This work introduces federated heterogeneous molecular learning. Federated learning allows end users to build a global model collaboratively while keeping their training data isolated. We first simulate a heterogeneous federated-learning benchmark (FedChem) by jointly performing scaffold splitting and latent Dirichlet allocation on existing datasets. Our results on FedChem show that significant learning challenges arise when working with heterogeneous molecules across clients. We then propose a method to alleviate the problem: Federated Learning by Instance reweighTing (FLIT(+)). FLIT(+) can align local training across clients. Experiments conducted on FedChem validate the advantages of this method. This work should enable a new type of collaboration for improving artificial intelligence (AI) in chemistry that mitigates concerns about sharing valuable chemical data.
format Online
Article
Text
id pubmed-9214329
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher Elsevier
record_format MEDLINE/PubMed
spelling pubmed-92143292022-06-23 Federated learning of molecular properties with graph neural networks in a heterogeneous setting Zhu, Wei Luo, Jiebo White, Andrew D. Patterns (N Y) Article Chemistry research has both high material and computational costs to conduct experiments. Intuitions are interested in differing classes of molecules, creating heterogeneous data that cannot be easily joined by conventional methods. This work introduces federated heterogeneous molecular learning. Federated learning allows end users to build a global model collaboratively while keeping their training data isolated. We first simulate a heterogeneous federated-learning benchmark (FedChem) by jointly performing scaffold splitting and latent Dirichlet allocation on existing datasets. Our results on FedChem show that significant learning challenges arise when working with heterogeneous molecules across clients. We then propose a method to alleviate the problem: Federated Learning by Instance reweighTing (FLIT(+)). FLIT(+) can align local training across clients. Experiments conducted on FedChem validate the advantages of this method. This work should enable a new type of collaboration for improving artificial intelligence (AI) in chemistry that mitigates concerns about sharing valuable chemical data. Elsevier 2022-06-02 /pmc/articles/PMC9214329/ /pubmed/35755872 http://dx.doi.org/10.1016/j.patter.2022.100521 Text en © 2022 The Authors https://creativecommons.org/licenses/by-nc-nd/4.0/This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
spellingShingle Article
Zhu, Wei
Luo, Jiebo
White, Andrew D.
Federated learning of molecular properties with graph neural networks in a heterogeneous setting
title Federated learning of molecular properties with graph neural networks in a heterogeneous setting
title_full Federated learning of molecular properties with graph neural networks in a heterogeneous setting
title_fullStr Federated learning of molecular properties with graph neural networks in a heterogeneous setting
title_full_unstemmed Federated learning of molecular properties with graph neural networks in a heterogeneous setting
title_short Federated learning of molecular properties with graph neural networks in a heterogeneous setting
title_sort federated learning of molecular properties with graph neural networks in a heterogeneous setting
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9214329/
https://www.ncbi.nlm.nih.gov/pubmed/35755872
http://dx.doi.org/10.1016/j.patter.2022.100521
work_keys_str_mv AT zhuwei federatedlearningofmolecularpropertieswithgraphneuralnetworksinaheterogeneoussetting
AT luojiebo federatedlearningofmolecularpropertieswithgraphneuralnetworksinaheterogeneoussetting
AT whiteandrewd federatedlearningofmolecularpropertieswithgraphneuralnetworksinaheterogeneoussetting