Andrew Song

The research group is dedicated to building next-generation AI tools for computational pathology, grounded in rigorous principles of statistical inference, with the overarching goal of deciphering multi-scale oncologic complexity. The lab's research will focus on developing state-of-the-art AI models (foundation models and/or AI agents) capable of integrating and cross-analyzing diverse data modalities—such as tissue images, spatial transcriptomics, spatial proteomics, and clinical reports—across multiple dimensions of clinical data (2D, 3D, and 4D longitudinal datasets). By combining these innovations with advanced statistical approaches such as Bayesian inference and optimal transport, the lab aims to open new frontiers in computational pathology and precision oncology.