Galaxy Formation and Evolution in the Data Science Era

Coordinators: Peter Behroozi, Francisco Villaescusa-Navarro, Shirley Ho, and Blakesley Burkhart

There is a vast discovery space in the application of astrostatistics and machine learning tools to galaxy formation and evolution. For example, current and future Integral Field Unit surveys are producing hundreds of spectra per galaxy across tens of thousands of galaxies, and galaxy morphology via imaging data contains a huge amount of information about the physical state of the system, down to the pixel level and across a range of wavelengths. Moreover, statistical and machine learning-powered outlier detection algorithms already find anomalous galaxies that do not fit into our current paradigm, and these detections will accelerate in the age of Rubin, DESI, Roman, Euclid, and the SKA. Data science tools will also be instrumental for linking observations with theoretical models, such as cosmological hydrodynamical simulations with resolved galaxy structure, or dark matter-only simulations and the semi-analytic or empirical models built on these.

This conference aims to explore the application of data-driven tools to learn about galaxy formation physics. It endeavors to maximize the gain from astrostatistics, data science, and machine learning for the galaxy formation field as a whole, by emphasizing the translation of data-driven results to physical understanding. This conference will focus on a sharing of expertise in data exploration and analysis tools, and an open discussion of how these may teach us about the physics of galaxy formation and evolution.