Open Sharing Icons

DHDP Tech Update: Flower AI Phase 1 complete

Flower by any other name; DHDP is budding in a secure and privacy-by-design garden for health data

What is DHDP? The Digital Health and Discovery Platform (DHDP) is a cloud-based ecosystem for precision medicine research that will share patient health care and clinical data securely across Canada. Ensuring security and privacy for this data is a priority. Once fully deployed, the Platform will help researchers, innovators, and entrepreneurs with data discovery, exploration and analysis by enabling AI and Machine Learning (ML).

What’s ‘under the hood’?  The Minimal Viable Product (MVP) stage — where platform functionality demonstrates proof of concept with synthetic data — was achieved in Winter 2025. In August, we will be providing a "sandbox” environment for invited users to test the Platform workflow using synthetic data and Federated Learning (FL) models.

Tech Update/What are we celebrating? Flower AI Phase 1 is complete. Implemented and evaluated by consultants, DT Consulting Group, DHDP has added the Flower AI framework into its ecosystem and shown that it not only works seamlessly beside current Platform architecture, but it also enhances functionality. Flower, described as offering ‘a unified approach to FL, analytics, and evaluation’, will be the backbone for the DHDP solution.

Why? Security by design Security is a priority for the DHDP, and FL is a key strategy for achieving this. Designed to boost FL, Flower brings flexibility and scalability to the Platform. The Flower AI ‘layer’ will help users work with diverse data sets that can be tricky simply because they’re a challenge to centralize. Flower will also help a wider range of users, from expert AI and ML data scientists to clinical researchers with basic programming skills, access to explore data.

What is Federated Learning (FL)?

FL helps AI and ML training but on-site, where the data is stored. This is important for privacy and security. The data that DHDP will handle is patient data — clinical notes, diagnostic information, medical imaging and genomic data — all identifiable and all requiring the highest levels of security. FL eliminates the need to move data around and gets rid of this potential for security breaches. Other benefits include minimizing loss during data transformation when it is moved.

 

How? Collaborative process Working with DT Consulting Group and Flower Labs, DHDP is developing proof of concept and premium features to create a functional backend for Platform users. Work continues on UI/UX front end interface development for users. Next steps include: robustness testing with QA/QC evaluation, Phase 2 full solution development and select users test-driving the latest version of Platform.

Building the core Flower will be the Platform’s core and DHDP’s three tech partners’ solutions are being integrated into this backbone.

“Think of each partner 'solution’ as a finished, fully furnished room within a house. We are building a house that connects and integrates these three rooms into a common look and feel for users,” explains DHDP Principal Architect and Strategy Lead, Ethan Hoang. “Flower is the foundation and structure of the house, and the rooms are accessed in a seamless and integrated fashion.”

 

Initially, the July sandbox environment gives workflow experience through tech partner, integrate.ai with Flower being brought on board for all DHDP tech partner solutions for early August.

Summary takeaway: Adding Flower AI to boost Federated Learning in the DHDP infrastructure has passed Phase 1 testing. Moving on to Phase 2 integration and then sandbox testing with invited users this summer.