NVidia launches Generative AI microservices for healthcare

NVidia has announced more than two dozen new microservices aimed at healthcare companies that want to exploit advances in Generative Artificial Intelligence (AI) from anywhere and from any cloud. These are optimised NVIDIA NIM™ AI models and workflows with standardised APIs that can serve as building blocks for creating and deploying cloud-native applications. They offer advanced imaging, natural language recognition, as well as predictive processes and digital simulation of biological systems.

In addition, NVidia accelerated development kits and tools, including Parabricks®, MONAI, NeMo™, Riva and Metropolis, can now be accessed as NVidia CUDA-X™ microservices to speed up workflows involved in drug discovery, medical image processing and genomic analysis.

According to NVidia, these new microservices could unlock countless opportunities for pharmaceutical companies, medical clinics and hospitals. They include the screening of drug compounds, the collection of patient data to help in the early detection of diseases and the implementation of smarter digital assistants.

Researchers and healthcare professionals can use these microservices to easily integrate AI capabilities into new or existing applications and run them from anywhere, in the cloud or locally, even adding copilot capabilities.

“For the first time, we can represent the world of biology and chemistry on a computer, making computer-aided drug discovery possible. By helping healthcare companies easily develop and manage solutions, we are enabling them to harness the full potential of Generative AI,” says Kimberly Powell, vice president of NVidia’s healthcare segment.

Inference resources

The new set of health microservices includes NVidia NIM, which provides optimised inference capabilities for a collection of imaging models, medical technologies, drug discovery and digital health. They can be used for prediction in the areas of biology, chemistry and molecular. NIM microservices are available via the NVidia AI Enterprise 5.0 software platform.

The microservices package also includes models for drug discovery, including MolMIM for generative chemistry, ESMFold for protein structure prediction and DiffDock to help understand how drug molecules interact with their targets. The VISTA 3D microservice speeds up the creation of 3D segmentation models. Universal DeepVariant offers 50 times greater speed in genomic analysis workflows compared to the basic DeepVariant implementation running on CPUs.

According to NVidia, around 50 application providers are already using healthcare microservices, as well as biotechnology and pharmaceutical companies and platforms, including Amgen, Astellas, DNA Nexus, Iambic Therapeutics, Recursion and Terray, and medical imaging software developers such as V7.

Microservices can also help improve direct patient care. Hippocratic AI is developing healthcare agents based on Generative AI with a focus on healthcare safety, which are connected to NVIDIA Avatar Cloud Engine microservices and use NVidia NIM for low-latency inference and speech recognition. These agents talk to patients over the phone to book appointments, carry out pre-operative activities, carry out post-discharge follow-ups, among other tasks.

Another company, Abridge, is developing an AI-based clinical conversation platform that generates draft notes, saving doctors up to three hours a day. According to the developer, going from raw audio in noisy environments to documents requires many AI technologies to work together seamlessly. Language identification, transcription, synchronisation and speaker diarisation must all take place within seconds. In addition, conversations must be structured according to the types of medical information contained in each utterance, and powerful language models must be applied to transform relevant evidence into summaries. Abridge claims that its system is capable of transforming clinical conversations into high-quality post-consultation documentation in real time.

You can try NVidia’s AI microservices at ai.nvidia.com and deploy NIM microservices via NVIDIA AI Enterprise 5.0 running on certified NVidia systems from vendors such as Dell Technologies, Hewlett Packard Enterprise, Lenovo and Supermicro, and on the main public cloud platforms such as Amazon Web Services (AWS), Google Cloud, Microsoft Azure and Oracle Cloud Infrastructure, as well as NVidia DGX™ Cloud.