Computer Vision / Video Analytics

Transforming AI Healthcare with Federated Learning

NVIDIA researchers, in collaboration with Owkin scientists, a premier member of NVIDIA Inception, as well as other scientists, this week published a new research paper on Nature Partner Journals Digital Medicine about the future of digital health with federated learning. 

“Existing medical data is not fully exploited by machine learning [ML] primarily because it sits in data silos and privacy concerns restrict access to this data,” the researchers stated in their paper.“ However, without access to sufficient data, ML will be prevented from reaching its full potential and, ultimately, from making the transition from research to clinical practice.”

As previously mentioned, Federated learning is a learning paradigm that allows developers and organizations to train a centralized deep neural network (DNN) with training data distributed across multiple locations. The technique makes it possible for organizations to collaborate on a shared model, without needing to directly share any clinical data. 

The new paper considers key factors that affect current healthcare AI applications and systems, including how key stakeholders such as patients, healthcare providers, hospitals, scientists, and AI developers are affected. 

“[Our work] explores how federated learning may provide a solution for the future of digital health, [including] the challenges and considerations that need to be addressed,” the researchers stated. 

Example federated learning workflows and difference to learning on a Centralised Data Lake. Source: npj Digital Medicine

NVIDIA announced the introduction of the first privacy-preserving federated learning system for medical image analysis in October 2019 at MICCAI. For that project, the team adapted a state-of-the-art training pipeline originally designed for data-centralized training and implemented it as part of the NVIDIA Clara Train SDK. For training and inference, the team used NVIDIA V100 Tensor Core GPUs

NVIDIA continues to work closely with academic and industry partners to develop federated learning applications for the scientific and AI community. 

Read the full Nature paper here

Also, see Owkin present in the NVIDIA Inception Premier Showcase October 7th online during GTC.

Discuss (0)