Speaker
Nicolas Cerardi
(EPFL)
Description
Machine Learning (ML) as emerged in the last decade as a powerful tool to solve complex, high-dimensional problems. This presentation will cover several aspects of my research related to ML techniques applied to tackle challenges in cosmology. Firstly, I will describe a new method to infer cosmological parameters from X-ray cluster number counts, using full-field emulation and simulation-based inference. Secondly, I will show how this framework can also be applied to radio interferometry, in order to constrain the HI fraction during the epoch of reionization (EoR). Lastly, I will present an alternative path to run cold dark matter simulations with Kolmogorov-Arnold Networks.
Primary author
Nicolas Cerardi
(EPFL)
