Technology 2 min read

NVIDIA Launches Federated Learning for Neural Networks

Called federated learning, this new approach will enable the training of a centralized healthcare deep neural network without jeopardizing clinical data.

Image courtesy of Shutterstock

Image courtesy of Shutterstock

In a breakthrough in healthcare AI studies, NVIDIA teamed up with researchers from King’s College London to develop the world’s first privacy-focused federated learning for deep neural networks.

In a press release, NVIDIA said that the collaboration is aimed at advancing medical research “while preserving data privacy and improving patient outcomes for brain tumor identification.”

To date, valuable data is needed to train deep neural networks (DNNs). It enables these intelligent systems to continuously learn and perform the tasks they are designed to do efficiently.

However, despite the many potential applications of AI in the medical field, developing and training healthcare AIs prove to be a challenging feat for researchers.

Unfortunately, the medical information needed to train healthcare DNNs, like patient data, is protected by privacy laws and policies. The latter significantly limits the size of datasets that AI scientists can use for their research.

But, that’s about to change with the new method developed by NVIDIA and King’s College researchers.

Privacy-Safe Federated Learning

NVIDIA described federated learning as a learning paradigm that enables the training of a centralized deep neural network using the information provided by organizations from different locations. With this approach, researchers from various healthcare organizations can collaborate without the need to share clinical data directly.

The researchers explained in their paper:

“Federated learning allows collaborative and decentralized training of neural networks without sharing the patient data. Each node trains its own local model and, periodically, submits it to a parameter server. The server accumulates and aggregates the individual contributions to yield a global model, which is then shared with all nodes.”

Federated learning already offers high levels of data security. But, to make it safer, the researchers also used the ε-differential privacy framework to determine privacy loss and protect patient and institutional data further.

In a statement to ZDNet, NVIDIA’s senior researcher, Nicola Rieke, said:

“We hope it will be a big step to enabling precision medicine on a large scale.”

Rieke and her team are scheduled to present their paper today at this year’s MICCAI world medical imaging conference in Shenzhen, China.

Read More: Scientists Use AI To Identify Brain Patterns Affecting Antidepressants

First AI Web Content Optimization Platform Just for Writers

Found this article interesting?

Let Rechelle Ann Fuertes know how much you appreciate this article by clicking the heart icon and by sharing this article on social media.


Profile Image

Rechelle Ann Fuertes

Rechelle is an SEO content producer, technical writer, researcher, social media manager, and visual artist. She enjoys traveling and spending time anywhere near the sea with family and friends.

Comments (0)
Most Recent most recent
You
share Scroll to top

Link Copied Successfully

Sign in

Sign in to access your personalized homepage, follow authors and topics you love, and clap for stories that matter to you.

Sign in with Google Sign in with Facebook

By using our site you agree to our privacy policy.