​​In healthcare, researchers and developers grapple with the daunting task of navigating extensive biological, chemical, and life science data to identify disease progression pathways, effective treatments, and potential drug candidates. Recently, foundation models—AI models trained on vast datasets using transformer architecture in a self-supervised manner—have emerged as transformative tools in this domain. Despite their promise in unlocking biological, chemical, and clinical insights, leveraging these models effectively remains challenging due to their inherent opacity, often described as "black boxes." To overcome this, we can design collaborative workflows that incorporate interactive visualizations to bridge the gap between AI models and experts. In this talk, I will present the latest approaches developed in collaboration with other researchers to aid users in interpreting model outputs and extracting meaningful insights. Furthermore, I will highlight examples demonstrating how these approaches have facilitated scientific discoveries.​

​​Bum Chul Kwon is a Staff Research Scientist on the Biomedical AI team at IBM Research. He also teaches graduate-level data visualization courses at Columbia University, the University of Pennsylvania, and the University of California, Berkeley. His research has been published in leading venues for visualization and human-computer interaction, including IEEE InfoVis, IEEE VAST, IEEE TVCG, and ACM SIGCHI. He has also contributed articles to prominent clinical journals such as Nature Communications, Diabetes, Scientific Reports, Cell Patterns, JMIR mHealth and uHealth, and BMC Medical Genomics. He serves on the program committees for IEEE InfoVis, PacificVis, ACM IUI, and the Visual Analytics in Healthcare Workshop. Previously, he was a postdoctoral researcher at the University of Konstanz. He holds a Ph.D. and M.S. in Industrial Engineering from Purdue University and a B.S. from the University of Virginia.​