Owkin, which is developing Federated Learning and AI technologies to advance medical research, announces it is teaming up with technology company NVIDIA and King’s College London (KCL) to deliver Federated Learning in the healthcare and life sciences sector. The King’s College London Medical Imaging and AI Centre for Value Based Healthcare (AI4VBH) is one of the world’s most ambitious Federated Learning projects in healthcare. It will initially connect four of London’s premier teaching hospitals before expanding throughout the UK, and will offer AI services to accelerate research and improve clinical practice in a wide range of therapeutic areas, including cancer, heart failure and neurodegenerative disease.
Owkin’s co-founder and Chief Scientific Officer, Gilles Wainrib, said: “This partnership brings together the best players in life science & healthcare, machine learning and data center infrastructure. NVIDIA’s platforms create the ideal and flexible footprint for hospitals to invest in machine learning. King’s College London has assembled the engineering, medical and data science talent, the high-quality patient data, and the governance framework in the AI4VBH Centre, which will show the world the future of healthcare analytics and the power of machine learning. Together we will be enabling the formation of a decentralized dataset that will generate enormous value for research and clinical practice.
“Owkin hopes to demonstrate that a Federating Learning architecture is safer for patients, and statistically equivalent to the traditional pooled model for analysis. Owkin also sees huge research potential to analyse the patient data in the AI4VBH Centre to identify new biomarkers, and high value subgroups for clinical trial design and diagnostics.”
KCL will use Owkin’s Federated Learning software and NVIDIA’s EGX Intelligent Edge Computing platform to develop research, clinical and operational improvements across a large number of clinical pathways, with cancer, heart failure, dementia and stroke likely areas of early focus. KCL will make this federated dataset available to its researchers as well as a community of Life Science & Healthcare companies that are consortium members within the Innovate UK funded AI Centre.
Sebastien Ourselin, Professor of Healthcare Engineering at KCL, said: “Our aim with the Innovate UK funded AI4VBH Centre is building a community of academic and private sector researchers that can collaborate to use advanced imaging and AI to develop insight as well as clinical and operational tools that will substantially improve the experience and the clinical outcomes of our patients.
“We are very pleased to welcome Owkin into our consortium of partners. Owkin are thought leaders in the new field of federated learning, and will make an important contribution to the AI Centre by providing the software layer that allows models to be built, orchestrated, secured and traced as they travel between our hospital and university partners. This is enables us to learn from data at scale, while preserving patient privacy. It also ensures that the predictive models developed from patient data are representative and unbiased because they will be trained on the widest possible population of patient data, which in phase 1 includes one third of the London population and will extend far beyond London in coming years. We truly see this architecture as the future of healthcare informatics”.
NVIDIA is the world leader in AI computing platforms that power the vast majority of machine learning infrastructure around the world. NVIDIA is enabling healthcare and life sciences to adopt the state of the art in AI and already provides many of the world’s leading hospitals, startups and instrument companies with their Clara Application Framework and a range of accelerated AI computing platforms, including the NVIDIA DGX and recently announced EGX platform.
Craig Rhodes, EMEA Industry Lead for Artificial Intelligence Healthcare and Life Science, NVIDIA said: “Owkin has developed a highly flexible, secure, traceable and privacy-preserving software stack that brings federated learning to life. It allows data scientists to design federated learning strategies and it orchestrates the travel of algorithms from one hospital to another. Importantly for the European context, where GDPR is a requirement, Owkin’s blockchain-based traceability platform gives hospital information governors the assurance they need that a patient’s data is being respected and protected.
“We are confident that the Owkin Connect stack, integrated with NVIDIA Clara and our EGX platform, is the right solution for KCL and many other leading research hospitals around the world. The collaboration with KCL, the Hospital Trusts, Owkin and NVIDIA brings together the clinical, data science and technology partners to ensure that patients can be confident in the integrity and security of their data in this new and rapidly evolving healthcare revolution.”
In the first phase, the AI4VBH Centre, which is part of KCL School of Biomedical Engineering, will federate the Data of three universities (King’s College London, Imperial College, and Queen Mary University of London) and four London Tier 1 teaching hospitals (Kings College Hospital, South London & Maudsley, Guy’s & St. Thomas’, and Barts Health). Phase 2 will extend the hospital network by up to 12 hospitals throughout the UK. The KCL School of Biomedical Engineering & Imaging Sciences recently won the Queen’s Anniversary Prize, which is the highest national honour awarded in the UK for higher and further education. Owkin are proud to be associated with this prize winning institution.
Owkin has integrated its Federated Learning Platform, Owkin Connect, with NVIDIA’s EGX, DGX and Clara platforms. NVIDIA’s EGX cloud-native, edge-first platform can deliver and activate a pre-integrated instance of Owkin Connect in a matter of minutes after plugging in an NGC-Ready for Edge Server. This is transformational for hospital research: Instead of building AI models on local data, hospitals will soon be able to extend their research to include data from the Owkin Loop – a network of Tier 1 Academic Medical Centres around the world, where scientists can train and validate models on decentralized data from behind the firewalls of participating hospitals. By expanding the size of the training dataset, Owkin ensures that models can be trained on heterogeneous data, without removing the data from the hospitals. The network already includes Tier 1 hospitals in France and the USA, and KCL and its London hospital partners will be the first participating academic medical consortium in the UK.