Following the talk on 25 September, “Clinical AI deployment made easy: Scale new solutions, in weeks not years”, this session will take a deeper dive into the underlying technology behind the Federated Learning Interoperability Platform (FLIP) built by Answer Digital for the NHS AI Centre for Valued Based Healthcare.
Come along if you’re a data scientist, AI model researcher or software engineer and discover how open-source solutions are being used to accelerate the development and deployment of AI models in the NHS.
Getting under the hood of FLIP, we’ll take a technical deep dive demonstrating how the platform is built and why it’s changing the way AI models are trained using healthcare data. We’ll cover how the OMOP common data model, secure data environments, NVIDIA FLARE open-source SDK, and MONAI collaborative frameworks are being combined to enable federated learning and deployment of AI applications.
During this talk, colleagues from the AI Centre, leading NHS trusts deploying the technology, and the solutions’ architects from Answer Digital will demonstrate:
– Opportunities for training early-stage models securely on large-scale, federated, clinical data sets
– How federated learning ensures no clinical data needs to leave an NHS Trusts network, increasing data privacy
– How FLIP shortens the validation window for AI models so they can be implemented in clinical pathways quicker
– Use cases being deployed in the NHS today
Attendees will also get a short demo of the AI Deployment Engine (AIDE) developed by the AI Centre for Value Based Healthcare. We’ll show how it allows models to be packaged up and easily deployed by enabling the testing of late-stage models and safe one-click deployment into a clinical workflow.
