Multi-source Data Integration for Precision Health and Disease Prediction

Currently, the field of healthcare is primarily focused on disease stages including treatments applied relatively late with suboptimal health outcomes. However, continued advancements in technology lead to the potential of detecting diseases early and enable preventative interventions. Prior AI models of time series data have been limited in scope, focusing on data from a single source (e.g., images or texts – free or structured data in the medical record) rather than incorporating data from multiple sources. I will develop methods for precision health, focusing on an efficient multi-source data integration of medical reports and radiology images. This kind of integration will fill the “information gap” of prior AI methods by providing comprehensive knowledge about the patient disease state (or even about a future disease from current healthy screening), and thus improve the prediction performance. I aim to fundamentally revolutionize healthcare leading to better and more productive lives for individuals.