Singapore’s National University Health System has built its AI production platform based on NVIDIA’s universal system for AI infrastructure.
According to a press statement, the NVIDIA DGX A100 runs at the core of the hospital group’s Endeavour AI platform to enable real-time predictions on diagnosis, progression of diseases, readmissions, risk of falls and others. It is running with five petaFLOPS (five quadrillion floating-point operations per second) of AI performance, making it possible to process various AI workloads in a single platform.
The system will also be integrated into the NUHS’ Discovery AI training platform to form a complete training and inference system.
WHY IT MATTERS
Dr Ngiam Kee Yuan, group chief technology officer at NUHS, said they needed NVIDIA GPUs for high speed and large volume inference processing. They would immediately run out of processing speed if they rely only on CPUs, he stressed.
NVIDIA’s technology also enables the AI tools in the Endeavour AI system to run quickly in the background to absorb data on a daily basis. “We are building a platform that enables multiple projects to run… Without the GPU, we cannot do a lot of these things,” Dr Ngiam explained. The Endeavour AI streams data and runs microservices that process all streaming data and produce outputs in real-time. It has the capacity to handle up to 150 projects, including those that involve structured medical data and text-based medical data for generating chatbots.
NVIDIA powering the Endeavour AI has led to an improvement in patient interactions with AI-powered chatbots, especially in appointment making; enhanced accuracy and speed of images, x-rays, scans processing for radiologists; and automated predictions “without even needing to click a button” while doctors are being alerted of at-risk patients. “These are tangible realities and outcomes that we expected when Endeavour AI went live,” Dr Ngiam said.
THE LARGER TREND
In a recent interview with Healthcare IT News, Dr Ngiam, who serves as the chief advisor to the NUHS Centre for Innovation in Healthcare, mentioned that in the next few years, they wanted to deploy and scale AI tools currently under research in clinical practice. After launching Discovery AI earlier, the NUHS rolled out its Endeavour AI platform.
Recently, the NUHS employed four applications from California-based enterprise data company TIBCO Software – BusinessWorks, Streaming, Messaging, and Spotfire – to support the integration of real-time medical data from EMR systems.
Speaking about personalised information, Dr Ngiam advised during a keynote session at the HIMSS21 APAC Conference, that all information needs to come together and not just be stored in silos that do not interact with each other.
ON THE RECORD
“There are many demands on healthcare these days and we are undertaking a digital transformation throughout the cluster. In the centre of our digital transformation is the use of AI. Advances in healthcare require great compute resources, and NVIDIA DGX A100 delivers easy access to performance needed to aid a world-class hospital,” Dr Ngiam shared.
“NVIDIA DGX A100 lets NUHS consolidate training, inference and analytics into a unified AI infrastructure. It will provide the computing power to help the hospital group achieve operational and scientific breakthroughs in the healthcare sector, benefitting clinicians and patients in Singapore,” said Dennis Ang, director of enterprise business for Southeast Asia and Australia-New Zealand Regions at NVIDIA.
Source: healthcareitnews.com
