Nvidia announced Monday it is open sourcing its FLARE software development kit to help healthcare providers collaborate on AI models.
FLARE stands for Federated Learning Application Runtime Environment and is the engine inside Nvidia Clara Train’s federated learning software used in AI applications for everything from medical imaging to genetic analysis and even COVID-19 research. With the SDK, data scientists can adapt existing machine learning work to a distributed approach.
Nvidia said it has worked with Roche Digital Pathology researchers and, separately, Erasmus Medical Center in the Netherlands on an application to identify genetic variants associated with cases of schizophrenia.
FLARE can work with open source MONAI for medical imaging as well.
In addition to the FLARE news, Nvidia said its AI Enterprise software is being used by the Netherlands Cancer Institute (NKI) to test AI workloads used to prep more precise 3D cancer scans. AI Enterprise has higher memory capacity for use with high-res images.
Also, Vyasa uses AI Enterprise to build apps to search unstructured content like patient care records, Nvidia said. The Nvidia announcement came at the annual meeting of the Radiological Society of North America.
“The hardest part of AI is not building AI, it’s getting data and annotating and having useful data,” said David Niewolny, director of business development for life sciences and healthcare at Nvidia, in comments with reporters. He estimated that 30% of all data on Earth is healthcare data.
“There’s tons out there but how can we make that usable?” he asked. “Data quality and prep are paramount, then how to keep it secure.”
Niewolny said Nvidia is “well-positioned to moving to democratizing AI and what customers want and need,” although he would not discuss Nvidia revenues specific to healthcare and life sciences.
“We’re in a unique position with medical devices going through a big transitional change…,” he added. “I’m exceptionally bullish on opportunities here. It’s going to be a big transition to the industry.”
In its most recent quarter, Nvidia reported $7.1 billion in record revenues, with $3.22 billion for gaming and $2.94 billion for data center. The professional visualization category at Nvidia was $577 million, up 144% from a year earlier. The company does not report a separate revenue line for healthcare and life sciences, but presumably some of those revenues are part of the data center and visualization categories.
RELATED: Nvidia broadens enterprise access to AI software with subscription plan