中央研究院 資訊科學研究所

Artificial intelligence (AI) and machine learning (ML) have become indispensable tools for translating complex biomedical and health data into actionable knowledge. This talk presents how explainable AI methods can be systematically integrated with computational biomedicine to enhance clinical decision support and population health modeling. Applications span from molecular-level bioinformatics, such as protein subcellular localization and allergen prediction, to patient-level analytics for in vitro fertilization outcomes, diabetic retinopathy screening, and preeclampsia prediction. Beyond the clinic, explainable ML approaches are also applied to public health informatics and infoepidemiology—demonstrated by predictive models for dengue fever outbreaks and COVID-19 surveillance using online behavioral data. Together, these studies illustrate how interpretable models not only improve transparency and trust in AI-driven healthcare but also bridge computational methods with human-centered medical decision-making. The talk concludes by outlining opportunities for interdisciplinary collaboration between biomedical data scientists and health informatics researchers, consistent with the vision of next-generation computational health programs.
Keywords: explainable AI, clinical decision support, computational biomedicine, health informatics, machine learning