PRS_Multi-Task-Learning

We developed a machine learning model to predict disease susceptibility using Polygenic Risk Scores (PRS) within a multi-task learning framework. I employed resampling techniques and utilized a shared encoder with a weights mechanism to learn both general and task-specific features. This approach demonstrated the effectiveness of combining medical data with advanced deep learning for disease risk prediction.