Galaxy Zoo in the Deep Learning Era - Mike Walmsley
Kavli Institute for Theoretical Physics via YouTube
Overview
Explore the evolution of Galaxy Zoo in the deep learning era through this 31-minute conference talk by Mike Walmsley from the University of Manchester. Delve into the vast potential of applying astrostatistics and machine learning tools to galaxy formation and evolution. Discover how current and future Integral Field Unit surveys are revolutionizing our understanding of galaxies by producing hundreds of spectra per galaxy across tens of thousands of galaxies. Learn about the wealth of information contained in galaxy morphology through imaging data, down to the pixel level and across various wavelengths. Understand how statistical and machine learning-powered outlier detection algorithms are uncovering anomalous galaxies that challenge our current paradigms, and how these discoveries will accelerate with upcoming projects like Rubin, DESI, Roman, Euclid, and the SKA. Explore the role of data science tools in bridging observations with theoretical models, including cosmological hydrodynamical simulations and dark matter-only simulations. Gain insights into the structure of Galaxy Zoo, self-supervised learning techniques, and the tools used in this citizen science project. Conclude with final thoughts on the impact of deep learning on galaxy classification and our understanding of the universe.
Syllabus
Introduction
Galaxy Zoo
Structure
Selfsupervised learning
Beyond Galaxy Zoo
Galaxy Zoo Tools
Final Thoughts
Taught by
Kavli Institute for Theoretical Physics