Computer Vision to Phenotype Human Diseases Across Physical and Molecular Scales
Explore cutting-edge computer vision algorithms for learning complex morphologies and phenotypes crucial to human diseases in this research seminar. Delve into examples spanning physical scales from macro to micro, including video-based AI for heart function assessment, spatial transcriptomics generation from histology images, and immune cell morphodynamics learning. Discover new design principles and tools for human-compatible and robust AI that enable these technologies. Gain insights from James Zou, an assistant professor at Stanford University and Chan-Zuckerberg investigator, as he discusses his work in developing novel machine learning algorithms to study human health and diseases, as well as making ML more reliable, accountable, and human-compatible.