Teaching Machines to Learn the Universe: Galaxy Science in the Rubin Era

Only a decade ago, projects like Galaxy Zoo relied on global human effort to classify millions of galaxies by eye. With the Vera C. Rubin Observatory’s first light now underway, these approaches no longer scale. Its Legacy Survey of Space and Time (LSST) will image tens of billions of galaxies, repeatedly mapping the sky over the next decade. Astrophysics has entered the big-data regime: deep learning algorithms are now required to not just process data but also identify the most scientifically meaningful signals. In this talk, we’ll explore how deep learning is reshaping galaxy science in the Rubin era. We’ll see how our models can help find unexpected objects, extract distances and physical properties from images alone, and uncover subtle structural patterns across cosmic time through accurate galaxy classification. Finally, we’ll look ahead to astronomy’s multi-modal future, where AI algorithms can integrate Rubin data with diverse surveys and data types to build a unified, data-driven picture of galaxy evolution.

Speaker: Michelle Park, Stanford University

Registration required to attend in person