Ji Young Byun
Ph.D. Candidate · Johns Hopkins BME
I develop methods to make AI systems more capable and reliable for healthcare and scientific discovery. My research spans large language models, vision-language models, and multi-agent systems, with a focus on improving reasoning and evaluating model capabilities and limitations. I’m a Ph.D. candidate at Johns Hopkins University, advised by Dr. Rama Chellappa, and expect to graduate in December 2026.
What I work on
- Reliable Systems
- Develop methods to improve reasoning in large language and vision-language models and identify when their answers can be trusted.
- Evaluation & Measurement
- Study how to evaluate model capabilities and limitations, including the reliability of LLM-based judges and whether evaluation metrics capture meaningful progress.
- Agentic Systems
- Build multi-agent systems that combine evidence, tools, and collaboration to tackle complex tasks, exploring what makes their reasoning and interactions more effective and reliable.
- Research to Practice
- Collaborate with researchers, clinicians, and engineers to connect advances in AI with practical needs in healthcare and science, aiming to build systems that support better decisions and scientific discovery.
Selected research
All Publications →- 2026
Overconfidence and Calibration in Medical VQA: Empirical Findings and Hallucination-Aware Mitigation
Byun, J. Y., Park, Y. J., Corbeil, J.-P., & Ben Abacha, A.
- 2026
Test-Time Scaling in Clinical Decision Making: An Empirical and Analytical Investigation
Byun, J. Y., Park, Y. J., Azizan, N., & Chellappa, R.
MIDL 2026 · 2026📄 paper - 2026
Adaptive Inference for Medical Vision Transformers: Token Reduction or Early Exit?
Byun, J. Y.*, Lee, H. S.*, Shuff, J. M., Venkatesh, R., Shekhawat, N. S., Parikh, K. S., & Chellappa, R.
- 2022
PAAN/MIF Nuclease Inhibition Prevents Neurodegeneration in Parkinson's Disease
Park, H., Kam, T. I., … Byun, J. Y., … & Dawson, V. L.
Cell · 2022📄 paper