Machine learning tests of the cosmological constant model and beyond – Savvas Nesseris

Speaker:  Savvas Nesseris (IFT, UAM-CSIC)

Abstract: We use a plethora of machine learning tools to test the cosmological constant model ΛCDM and probe for covariant modifications of general relativity (GR). In particular, we use genetic algorithms (a type of evolutionary symbolic regression algorithm) to first test the swampland conjectures and second, to reconstruct the Weyl potential, which is a probe of modifications of GR, such as f(R) theories. In a different vein, we also use neural networks to perform model selection between f(R) models and ΛCDM, and we compare with the traditional statistical analyses. In all cases we highlight the interpretability of our methodology and how it impacts current and future large scale structure surveys.