Sihang Zeng

Sihang Zeng is a PhD student in Biomedical and Health Informatics at University of Washington. Previously he received bachelor’s degree in Electronics Engineering with a minor in Statistics at Tsinghua University in China. Sihang’s research interests lie in the intersection of deep learning and biomedical informatics. His research mainly focuses on representation learning and large language models, including biomedical term representations, protein and variant representations, and patient representations, with applications including term clustering and rare disease diagnostics. Sihang developed a fine-grained term representation model CODER++, which achieved best performance in the task of term clustering and was used to construct a large-scale biomedical knowledge graph BIOS. Sihang is currently focusing on patient representation using temporal EHR data, with the application in prostate cancer survival analysis, and the broad use of large language models in biomedical domain.