Graph representation learning
Learning from relationships, not just individual observations. I study graph-based and semi-supervised methods that use network structure when labels are scarce.
Semi-supervised learning
Graph learning × Biomedical AI
Learning fromconnections.
Sunghong Park, Ph.D.
I develop machine learning methods that connect heterogeneous data, learn with limited labels, and bring biological structure into biomedical AI.
From data to understandingFIG. 01
Learning useful representations from connected observations.
Conceptual research map · not experimental data13Peer-reviewed
journal articles
journal articles
17Patent applications
KR · PCT · US
KR · PCT · US
02NRF research grants
as Principal Investigator
as Principal Investigator
Postdoctoral Researcher · KAISTLife Science Research Institute
Research directionsStructure. Transfer.
Structure. Transfer.
Understanding.
Three connected directions for learning from complex biomedical data.
Explainable biomedical AI
Connecting neuroimaging, plasma proteomics, and biological knowledge to study disease heterogeneity and interpretable biomarkers.
Multimodal integration
Learning across domains & time
Transferring information between heterogeneous populations, measurement settings, and time points while preserving useful differences.
Domain adaptation
Selected publications
All 13 journal articlesIdeas into evidence.
Explainable multiplex graph propagation network with multimodal neuroimage integration for dementia subtype diagnosis
Integrating multiple neuroimaging modalities through an explainable multiplex graph propagation network for dementia subtype diagnosis.
NeuroFANN: identification of neuropathological subtypes in dementia with plasma proteins by using functionally annotated neural network
Using functional annotations and plasma protein relationships to study neuropathological subtypes of dementia.
Prospective Domain Adaptation for Longitudinal Data
Adapting representations across observation times to study prediction with longitudinal data.
Published work only. Manuscripts under review are listed separately in the publication archive.
Currently exploring
Research projectsThe next questions.
Explainable AI for long-term brain age trajectories
Connecting multimodal diagnostic estimation with long-term prognostic prediction. An NRF-supported investigation into explainable brain aging models.
Explore the project