Bridging Theory and Data: Correcting Nuclear Mass Models with Interpretable Machine Learning | ResearchPod