Federated computing platform for collaborative global health innovation and societal impact
Federated Computing in Open Health
Open Health is built on federated computing—a way for organisations to work with data together without moving or sharing the data itself. Instead of centralising sensitive data, computation is sent to where data already lives, and only approved results are returned.
​
To make this work in real health systems, Open Health separates three complementary layers: Federated Learning (FL), Federated Computing (FC), and Federated Computing as Code (FCaC).
Federated Learning (FL): Training AI models without sharing data
​Federated Learning enables multiple organisations to train a shared AI model while keeping data local. Each partner trains the model on its own data and shares only model updates, which are securely aggregated.
​
What FL is good for:
-
Improving prediction models (e.g. disease risk, outcomes)
-
Situations where the question and model are known in advance
​
Limitations:
-
FL is model-driven and narrow in scope
-
It does not handle governance, permissioning, or broader data exploration
Federated Computing (FC): Running analytics and exploration across organisations
Federated Computing goes beyond model training. It supports a wider range of federated workloads, including analytics, inference, statistical analysis, cohort building, and data processing—executed locally at each organisation.
​
What FC enables:
-
Exploring distributed data before committing to a specific model
-
Iterative, real-world decision making across organisations
-
Collaboration when questions evolve over time
​
This makes FC essential for health systems, where understanding populations, contexts, and trade-offs often comes before building an AI model.
Federated Computing as Code (FCaC): Governance built into the system
FCaC is the layer that makes federated collaboration governable at scale. Instead of relying on manual processes or platform-specific logic, rules about who can run what, where, and under which conditions are encoded directly into the system.
​
What FCaC provides:
-
Automated admission and boundary enforcement
-
Clear, auditable evidence that rules were followed
-
Trust and transparency across organisations
​
FCaC does not judge whether an analysis is “good” or “ethical”—it ensures that only approved workloads run under declared constraints.
Why this matters for Open Health
Together, FL, FC, and FCaC enable Open Health to support real, multi-partner collaboration across universities, healthcare providers, public bodies, and industry. This separation allows organisations to explore data, build AI solutions, and generate shared insight—while maintaining data sovereignty, regulatory compliance, and institutional autonomy.
​
This is what makes Open Health scalable, trustworthy, and deployable at national and global levels.
​