Hong-Jun is a research scientist who has interests broadly in the areas of accelerated training, validation, and optimization in high-performance computing environments. He is involved in the Joint Design of Advanced Computing Solutions for Cancer (JDACS4C) and has developed deep learning-based information extraction models from free-form clinical texts. He is also involved in the Exascale Computing Project, Cancer Distributed Learning Environments (ECP-CANDLE). He is developing scalable training solutions for the information extraction models and researching enhancing scalability and clinical task performance by applying ensemble, model distillation, and compression. He led research on scalable artificial intelligence models for high throughput multi-labeled medical image analysis and several medical imaging projects.
He has experience in cancer epidemiology, population-level health data analytics, and modeling and simulation. He also has experience in observer perception and performance studies with eye-tracking apparatus.
Before joining the lab, he was a research staff at the University of Pittsburgh Medical Center, developed statistical analysis tools and simulations in high-performance computing environments.
Last Updated: February 25, 2021 - 12:05 pm