Collaboration

I welcome collaboration opportunities and encourage you to reach out if our research interests align. My current research directions include:

• Precision Healthcare with Multimodal Electronic Health Record (MMEHR) Data: Multimodal EHR modeling presents several open challenges, including extracting meaningful information from unstructured and sparse clinical notes, modeling irregularly sampled longitudinal data such as laboratory tests, and bridging the gap between patient-level EHR data and relevant biomedical knowledge. I am interested in developing reliable and interpretable methods that integrate heterogeneous clinical data for risk prediction and precision healthcare.

• Drug Repurposing and Discovery (DRD): Modern drug discovery increasingly leverages professionally curated biomedical knowledge graphs, multi-omics data, and the reasoning capabilities of large language models (LLMs). However, important challenges remain in computational efficiency, reliability and reproducibility of generated hypotheses, biological interpretability, and effective integration of multi-omics evidence. My research aims to develop AI methods for reliable and interpretable therapeutic hypothesis generation.

• Fundamental Algorithms for Graph Learning and Transfer Learning: Graph learning and transfer learning provide promising foundations for addressing heterogeneous data integration, limited high-quality labeled data, and distribution shifts across domains and institutions. I am interested in developing novel learning algorithms that address these fundamental challenges while supporting biomedical applications. My current work includes transfer learning for multi-center EHR modeling and context-conditioned graph learning for reliable biomedical knowledge discovery.