Special Physics Seminar

A data-driven approach to 21cm cosmology

Adrian Liu

University of California, Berkeley

Current and upcoming radio telescopes such as the Canadian Hydrogen Intensity Mapping Experiment (CHIME) and the Hydrogen Epoch of Reionization Array (HERA) promise to survey an unprecedented portion of our observable Universe, the latter for example covering ~200 cubic Gpc of comoving volume. Such efforts represent the general recent trend of radio cosmology becoming a data-rich enterprise. In this talk, I will discuss recent efforts in embracing a data-driven approach to radio observations. This includes a discussion of the extended Global Sky Model project, whose goal is to use statistical methods to ask the most fundamental question of survey science: “What does the sky look like in all directions at all wavelengths?” Specializing to 21cm observations, I will then discuss how machine learning techniques make it practical to fully explore the parameter space of astrophysical and cosmological parameters spanned by simulations of Cosmic Dawn, the epoch when the first stars and galaxies formed. I will conclude with an overview of how such techniques will be used by HERA to place qualitatively new constraints on Cosmic Dawn and the subsequent Epoch of Reionization, as well as on fundamental physical�parameters such as the sum of the neutrino masses.