NAIRR pilot expands research access to NVIDIA AI infrastructure

The U.S. National Science Foundation’s National Artificial Intelligence Research Resource (NAIRR) pilot has supported over 700 projects across areas including protein prediction and infectious disease outbreak management. NVIDIA contributed a cloud-based resource that gives researchers dedicated access to a minimum of four NVIDIA DGX nodes for at least a month, along with technical support for onboarding and project execution.

Several participating teams are applying the infrastructure to scientific foundation models. Polymathic AI, a coalition involving the Flatiron Institute, Cambridge University and Lawrence Berkeley National Lab, is using NVIDIA GPUs and NVIDIA NVLink interconnect technology to strengthen fluidlike physical simulations through its Well dataset and Walrus foundation model. University of Michigan researchers are building MIST, a family of molecular foundation models for exploring chemical space and energy materials, pretrained on large unlabeled molecular datasets and fine-tuned on more than 400 structure-property relationships.

MIST was developed on a 40-GPU NVIDIA DGX cluster from a NAIRR allocation and an additional 200,000 NVIDIA GPU hours on ALCF’s Polaris cluster. Boston University is also using NVIDIA accelerated compute for BEACON, an outbreak monitoring pipeline that analyzes signals from sources such as HealthMap, news, social media and expert communications. According to Ioannis Paschalidis, reports that previously took several hours can now be produced in roughly two minutes.