Speaker
Description
With ongoing and future experiments, we are set to enter a more data-driven era in astronomy and astrophysics, for example with interferometric measurements of the 21-cm signal but also with observations in the far-infrared, optical, UV, and beyond. Both larger-scale techniques such as multi-line intensity mapping and higher sensitivity surveys warrant the need for efficient data reduction and automation as well as the ability to extract more and less biased information. To optimally learn the Universe from low to high redshift I advocate for new observational techniques such as multi-line intensity mapping as well as the application of modern machine learning techniques. In 3D, tomography of line intensity maps such as the 21-cm line of hydrogen targeted by the Square Kilometre Array (SKA) can teach us about properties of sources, gaseous media between and cosmological structure formation. I showcase the use of deep networks that are tailored for the 3D structure of tomographic 21cm light-cones of reionisation and cosmic dawn to directly infer e.g. dark matter and astrophysical properties jointly without an underlying Gaussian assumption. This high-redshift study is complemented with recent lower redshift machine learning results for the SKA data challenge, where our team detected and characterised sources in a large TB data-cube of the hydrogen 21cm line.
