NVIDIA releases generative AI microservices to promote drug research and development, medical technology and digital health development

New NVIDIA NIM and GPU-accelerated microservices developed for biological, chemical, imaging and medical data and run in the NVIDIA DGX cloud

SAN JOSE, Calif., USA - GTC - March 18, 2024 PT - NVIDIA today launched more than two dozen new microservices to enable global healthcare enterprises to take advantage of the latest advances in generative AI, anywhere and on any cloud. .

The new NVIDIA Healthcare Microservices suite includes optimized NVIDIA NIM™ AI models and workflows, and provides industry-standard application programming interfaces (APIs) for creating and deploying cloud-native applications. They provide advanced medical imaging, natural language and speech recognition, and digital biology generation, prediction, and simulation capabilities.

In addition, NVIDIA accelerated software development suites and tools, including Parabricks®, MONAI, NeMo™, Riva, Metropolis, are now accessible through NVIDIA CUDA-X™ microservices to accelerate medical research and development, medical imaging, genomics analysis and more Workflow.

These microservices, 25 of which are launching today, can accelerate the transformation of healthcare enterprises as generative AI opens up numerous opportunities for pharmaceutical companies, doctors and hospitals. These include screening trillions of pharmaceutical compounds to advance medicine, collecting more complete patient data to improve early disease detection, enabling smarter digital assistants, and more.

Researchers, developers, and healthcare practitioners use these microservices to easily integrate AI into new and existing applications and run them anywhere from the cloud to on-premises, enhancing the life-saving work they do. Work.

"For the first time ever, we are expressing the world of biology and chemistry in computers, making computer-aided drug discovery possible," said Kimberly Powell, vice president of NVIDIA's medical business . "With our help, medical companies can easily build and manage AI solutions. solutions to harness the full power and potential of generative AI.”

NVIDIA NIM Healthcare Microservices for Inference

NVIDIA NIM in the new medical microservices suite provides optimized inference for the growing number of models in medical imaging, medical technology, drug discovery and digital health. These models can be used in generative biology and chemistry as well as molecular prediction work. NIM microservices are now available through the NVIDIA AI Enterprise 5.0 software platform.

This set of microservices also includes a series of models for drug development, such as the generative chemistry model MolMIM, the protein structure prediction model ESMFold, and the model DiffDock that helps researchers understand how drug molecules interact with targets. VISTA 3D microservices accelerate the creation of 3D segmentation models. Universal DeepVariant microservices can speed up variant identification in genome analysis workflows by more than 50 times compared to ordinary DeepVariant running on the CPU.

Cadence, a leading computational software company, is integrating NVIDIA BioNeMo microservices for AI-guided molecular discovery and lead optimization into its Orion® molecular design platform for accelerating drug discovery.

With Orion, researchers at pharmaceutical companies can generate, search and model databases of hundreds of billions of compounds. BioNeMo microservices such as MolMIM generative chemical model and AlphaFold-2 protein folding model have greatly enhanced Orion's design capabilities.

"Our pharmaceutical and biotech customers need access to accelerated resources for molecular simulations," said Anthony Nicholls, corporate vice president at Cadence . "By leveraging BioNeMo microservices, researchers can generate molecules optimized for the scientists' specific needs."

Nearly 50 application providers and multiple biotech and pharmaceutical companies and platforms are using medical microservices, including Amgen, Astellas, DNA Nexus, Iambic Therapeutics, Recursion and Terray, as well as medical imaging software manufacturers such as V7.

David M. Reese, executive vice president and chief technology officer at Amgen, said: "Generative AI is transforming drug development, allowing us to build advanced models and seamlessly integrate AI into the antibody design process. Our team is leveraging this Technology develops next-generation medicines that will deliver the greatest value to patients.”

Improve patient-clinician interactions

Generative AI is changing the future of patient care. Hippocratic AI is developing task-specific generative AI medical agents powered by the company’s safety-focused medical Large Language Model (LLM), connected to NVIDIA Avatar Cloud Engine (ACE) microservices and will use NVIDIA NIM for low-level Delayed reasoning and speech recognition.

These AI medical agents talk to patients and complete appointments, pre-operative reminders, and post-discharge follow-up.

Munjal Shah, co-founder and CEO of Hippocratic AI , said: “With generative AI, we can address some of the healthcare industry’s most pressing needs, helping alleviate widespread staffing shortages and provide more high-quality care while improving Patient outcomes. NVIDIA’s technology stack is critical to enabling the conversational speed and fluidity that are integral to creating natural, emotional connections between patients and Hippocratic’s generative AI medical agents.”

Abridge is building an AI-driven clinical conversation platform. The platform generates draft clinical notes, saving clinicians up to three hours per day. Converting raw audio in a noisy environment into a document requires seamless collaboration between multiple AI technologies: speech recognition, transcription, correction and speaker recognition (diarization) must all be completed within seconds, and the dialogue must be based on The kinds of medical information contained in each sentence are constructed, and powerful language models need to be applied to transform relevant evidence into summaries. The system converts clinical conversations into high-quality post-diagnosis documentation in real time.

Models created by Flywheel can be converted into microservices. The company's centralized, cloud-based platform empowers biopharmaceutical companies, life sciences organizations, medical providers and specialty medical centers to identify, organize and train medical imaging data to reduce time to insight.

Trent Norris, chief product officer of Flywheel , said: "Today, with the rapid development of medical technology, the integration of NVIDIA's generative AI microservices and the Flywheel platform is a transformative leap. With these advanced tools, we not only enhance our capabilities in medical imaging and Data management capabilities also greatly accelerate medical research and patient care outcomes. Powered by cutting-edge AI solutions built by NVIDIA, Flywheel's AI Factory can meet the needs of medical customers anytime, anywhere, helping to advance the fields of digital health and biopharmaceuticals development of."

Availability

Developers can visit ai.nvidia.com to try out NVIDIA AI microservices and deploy production-grade NIM microservices with NVIDIA AI Enterprise 5.0 running on NVIDIA certified systems from providers including Dell Technologies, HPE, Lenovo and Supermicro. and leading public cloud platforms such as AWS, Google Cloud, Microsoft Azure and Oracle Cloud Infrastructure. Additionally, it is available for trial on the NVIDIA DGX Cloud.

For more information, please visit the GTC NVIDIA booth online from March 18 to 21, and watch the replay of NVIDIA founder and CEO Jensen Huang’s keynote speech.