Industry-Scale Orchestrated Federated Learning for Drug Discovery
Abstract
To apply federated learning to drug discovery we developed a novel platform in the context of European Innovative Medicines Initiative (IMI) project MELLODDY (grant n°831472), which was comprised of 10 pharmaceutical companies, academic research labs, large industrial companies and startups. The MELLODDY platform was the first industry-scale platform to enable the creation of a global federated model for drug discovery without sharing the confidential data sets of the individual partners. The federated model was trained on the platform by aggregating the gradients of all contributing partners in a cryptographic, secure way following each training iteration. The platform was deployed on an Amazon Web Services (AWS) multi-account architecture running Kubernetes clusters in private subnets. Organisationally, the roles of the different partners were codified as different rights and permissions on the platform and administrated in a decentralized way. The MELLODDY platform generated new scientific discoveries which are described in a companion paper.
Cite
Text
Oldenhof et al. "Industry-Scale Orchestrated Federated Learning for Drug Discovery." AAAI Conference on Artificial Intelligence, 2023. doi:10.1609/AAAI.V37I13.26847Markdown
[Oldenhof et al. "Industry-Scale Orchestrated Federated Learning for Drug Discovery." AAAI Conference on Artificial Intelligence, 2023.](https://mlanthology.org/aaai/2023/oldenhof2023aaai-industry/) doi:10.1609/AAAI.V37I13.26847BibTeX
@inproceedings{oldenhof2023aaai-industry,
title = {{Industry-Scale Orchestrated Federated Learning for Drug Discovery}},
author = {Oldenhof, Martijn and Ács, Gergely and Pejó, Balázs and Schuffenhauer, Ansgar and Holway, Nicholas and Sturm, Noé and Dieckmann, Arne and Fortmeier, Oliver and Boniface, Eric and Mayer, Clément and Gohier, Arnaud and Schmidtke, Peter and Niwayama, Ritsuya and Kopecky, Dieter and Mervin, Lewis H. and Rathi, Prakash Chandra and Friedrich, Lukas and Formanek, András and Antal, Peter and Rahaman, Jordon and Zalewski, Adam and Heyndrickx, Wouter and Oluoch, Ezron and Stößel, Manuel and Vanco, Michal and Endico, David and Gelus, Fabien and de Boisfossé, Thaïs and Darbier, Adrien and Nicollet, Ashley and Blottière, Matthieu and Telenczuk, Maria and Nguyen, Van Tien and Martinez, Thibaud and Boillet, Camille and Moutet, Kelvin and Picosson, Alexandre and Gasser, Aurélien and Djafar, Inal and Simon, Antoine and Arany, Adam and Simm, Jaak and Moreau, Yves and Engkvist, Ola and Ceulemans, Hugo and Marini, Camille and Galtier, Mathieu},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2023},
pages = {15576-15584},
doi = {10.1609/AAAI.V37I13.26847},
url = {https://mlanthology.org/aaai/2023/oldenhof2023aaai-industry/}
}