Published September 27, 2024 | Version v1
Project deliverable Open

Deliverable 2.14: Computing infrastructures

  • 1. SURF
  • 2. ROR icon Aalborg University
  • 3. ROR icon University of Padua

Description

One of the main goals of the HEREDITARY project is establishing secure and scalable computing infrastructures in local medical centres to support federated learning. These infrastructures enable data processing and collaboration without centralising sensitive data. These infrastructures include managed services, computational resources, and storage facilities, ensuring medical centres can actively contribute to the federated learning process while maintaining data privacy.

This deliverable's milestone is setting up the necessary computing and storage infrastructures in medical centres, specifically at UCD, RUMC, UNITO, and UNIPD, and testing the communication protocol that connects these centres to the federated learning infrastructure. This deliverable will show that each medical centre is advancing toward meeting these infrastructure goals and that initial communication protocols have been tested and verified.

Task 2.6 aims to implement the federated learning infrastructure. However, in the context of Deliverable D2.14, the critical objective is to establish computing infrastructures in UCD, RUMC, UNITO, and UNIPD, ensuring that they have the necessary facilities to host federated learning components, process their data locally, and communicate securely with other partners.

We are reporting on the progress in setting up computing infrastructures at UCD, RUMC, UNITO, and UNIPD, including:

• UCD, RUMC, and UNIPD have the required infrastructure in place.
• UNITO is in the process of acquiring the necessary infrastructure.
• Initial communication protocols have been tested between UCD, HES-SO,
RUMC, SURF, and UNIPD. Once their infrastructure is ready, further expansion 
is planned for full collaboration with UNITO.

Additionally, the HES-SO workshop conducted earlier this year successfully demonstrated communication between these medical centres, providing evidence of networking and data-sharing capabilities in a federated learning environment.

Files

Hereditary_D2.14_V1.5.pdf

Files (527.3 kB)

Name Size Download all
md5:9fa06eb38db8b47cc3991464fd474bfc
527.3 kB Preview Download

Additional details

Funding

HEREDITARY – HetERogeneous sEmantic Data integratIon for the guT-bRain interplaY 101137074
European Commission