University of Florida researchers joined forces with scientists at NVIDIA, UF’s partner in its artificial intelligence initiative, and the OpenACC organization to significantly accelerate brain science as part of the Georgia Tech GPU Hackathon held last month. The hackathon is designed to help computational scientists and researchers optimize their applications.
During the hackathon, the UF-NVIDIA team worked to improve processing speed for a UF study that uses artificial intelligence technology to develop precision dosing for a non-pharmaceutical treatment aimed at preventing dementia. Using UF’s powerful supercomputer, the Hipergator, the team drastically improved the processing time for evaluating tens of thousands of datapoints collected from an individual brain. At the outset of the hackathon, the team’s algorithm took nearly 15 hours to analyze the data from one brain. After optimizing the code throughout the hackathon, the processing speed is now more than 40 times faster, dropping the run time to just half an hour per brain.
The team was composed of UF members Adam Woods, Ph.D., an associate professor of clinical and health psychology in the College of Public Health and Health Professions and associate director of the Center for Cognitive Aging and Memory at UF’s McKnight Brain Institute; Ruogu Fang, Ph.D., an assistant professor of biomedical engineering at the College of Engineering; Aprinda Indahlastari, Ph.D., a research assistant professor of clinical and health psychology; and Alejandro Albizu, a doctoral student in neuroscience at the College of Medicine. NVIDIA team members included Oded Green, Ph.D., a solutions architect, Kaleb Smith, Ph.D., a senior data scientist, and Jingchao Zhang, Ph.D., NVIDIA’s resident scientist at the UF NVIDIA AI Technology Center.