The project ODELIA, financed under Horizon Europe, builds the first pan-European swarm learning (SL) network that allows for privacy conserving training of medical artificial intelligence (AI) algorithms with true democratic participation of all partners. With the help of such network, AI algorithms will be developed and validated for detection of breast cancer in magnetic resonance imaging (MRI) screening examinations as a demonstration case. Training clinically useful AI models usually requires sharing of patient related data which often faces several obstacles. This problem has been tackled by federated learning (FL). However, in FL AI models are combined by a central coordinator requiring participants to relinquish control over the AI model, concentrating control and commercial exploitation in a single actor. This limitation has been addressed by swarm learning (SL), in which AI models are trained decentrally and models are combined without the requirement for a central coordinator. SL in medical AI has never been applied in a real-world large-scale setting before, which is the main objective of ODELIA.
In ODELIA, we will build a pan-European academic and clinical consortium to develop, implement and evaluate SL-based workflows to train AI models in medical imaging, in particular- and as an exemplary demonstration case in the context of breast cancer screening. We will create a blockchain that enables institutions across Europe to jointly train AI models without having to share their data or rely on a central coordinator.
ODELIA includes three types of technologies. Firstly, the project includes an AI algorithm which learns to detect cancer tissue in MRI data. Secondly, an SL algorithm will be developed, which enables multiple partners to jointly train AI algorithms without sharing any data. Lastly, an online viewer will be used which allows anyone to apply swarm-learning-trained AI algorithms on their own data.
Facts and figures
Coordinator: European Institute for Biomedical Imaging Research (EIBIR)
Number of Partners: 12
Start Date: January 1, 2023
End Date: December 31, 2027
Total Funding: € 8,691,755.00
This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No. 101057091.
Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Health and Digital Executive Agency (HADEA). Neither the European Union nor the granting authority can be held responsible for them.