Governance & Organisation
Governing bodies
The organizational structure has been established specifically to handle the size and complexity of the Project by ensuring clear reporting lines, effective governance and decision making and sufficient advisory input. It reflects the requirements for successful close collaboration between the public and industry partners.
Work Packages organisation
The Project is structured into 7 distinct Work Packages that address specific Project objectives but are clearly interdependent in achieving the overall milestones and deliverables.
Work Package 1
Preprocessing of the data to a level of necessary and sufficient standardization
Work Package 2
Federated and privacy-preserving machine learning algorithms
Work Package 3
Arbitration across algorithmic options and evaluation of predictive gain in platform run
Work Package 4
Implementation of enterprise-enabled (i.e. audit-ready) software
Work Package 5
Secure infrastructure and software deployment and industrial IT-technical scoping
Workgroup 6
Platform operation and monitoring and sustainability and service-related dissemination
Workgroup 7
Overall Project management and communication
Project Roadmap
Platform development will be orchestrated around yearly execution runs. During the 36 months of the Project there will be 3 iterations of specification, development, test and evaluation, deployment to production and run.
Evaluation and testing runs will be organized each year to assess the performance of the proposed approach, both for the platform security and the relevance of federated learning applied to Drug Discovery. Academic research in ML privacy, cryptographic security, Drug Discovery and multi-task federated learning will be continuous throughout the Project and will feed the specifications and features to be integrated into the platform.