Oracle
3072 Case Studies
A Oracle Case Study
Emory University’s ViTAL Lab faced a challenge building machine learning models to measure Parkinson’s disease symptoms from video and wearable data, which required heavy GPU computing and memory. Its on-premises system was too limited and busy to support the research, so the team turned to Oracle Cloud and Oracle’s OCI Compute and GPU instances.
Oracle implemented HIPAA-compliant cloud compute with bare metal GPUs, enabling Emory University to run 4 to 8 GPUs in parallel and avoid VM overhead. This cut experiment time from four days to just a few hours, and the resulting model matched expert freezing-of-gait ratings with 96.2% precision, significantly accelerating the team’s Parkinson’s research.
Hyeokhyen Kwon
Assistant Professor, Biomedical Informatics