TYPE OF SUPPORT
Research Background
Australia is developing new ways to safely analyse health data from different locations, without needing to move the data around.
This approach, called Federated Learning (FL), lets hospitals, research groups, and registries work together to train models while keeping their data secure and local.
This project brings together national partners to make it easier to choose the right tools, manage risks, and provide training, so researchers can use this method confidently across different states and organisations.
QCIF Role
QCIF leads two core streams and contributes technical and training expertise across the program:
FLAVRE: Lead the design of a Reference Secure Virtual Research Environments Architecture for Federated Learning (FL) and Federated Analysis (Nectar‑deployable, ISO27001‑aligned), including authentication, access control (e.g., attribute‑based encryption), monitoring, backup and data protection.
Adoption & Skills: Lead development of FL training materials (software/platform recommendations; node/server/network implementation; governance and trust frameworks) and deliver recorded training and early‑adopter support.
Leadership & governance: QCIF provides leadership, governance and oversight on assigned work packages via Work Package Lead and Steering Committee roles.
Technical contribution: QCIF technology leads provide guidance on FL software and architecture; QCIF skills coordination supports design and delivery of training and engagement.
Research Outcome & Impact
By project completion, Australia will have a recommended pathway to select Federated Learning platforms; a reference secure architecture (FLAVRE) suitable for health; evidence‑based risk, standards and governance frameworks; and training packages for expert users.
This reduces time‑to‑initiate FL projects, increases trust and interoperability across nodes and networks, and accelerates health research using real‑world data.
By project completion, Australia will be equipped with trusted frameworks, secure architecture, and expert training to accelerate federated learning in health, reducing project initiation time, enhancing interoperability, and unlocking the power of real-world data for research.
Dr Moji Ghadimi, Head of AI & Quantum Algorithms, QCIF
Collaborating Organisations
UNSW (Lead; ACDN) | University of Sydney (AIS/NIF) | University of Queensland (NINA) | ARDC
