Speaker
Description
This work focuses on the growth of cosmic structures within scale-dependent modified gravity frameworks, using the Hu-Sawicki model as a working example. In these scenarios, the effective gravitational coupling, Geff, depends on the scale. Consequently, the matter perturbation equation must be solved individually for every Fourier mode k, which inherently requires scale-dependent initial conditions. This scale dependence is crucial and, in general, cannot be neglected (not even in the sub-horizon approximation). Furthermore, obtaining these conditions through standard numerical integration can carry a prohibitively high computational cost, limiting the viability of cosmological parameter inference. To circumvent this computational issue, we present an artificial Neural Network (NN) trained to continuously and accurately predict these initial conditions across a wide parameter space in a cuasi-instantaneous way. We compute the matter perturbations and their derivatives using this trained NN to compare our theoretical predictions with current observational data. In doing so, we highlight and discuss the conceptual and methodological challenges that arise when testing modified gravity models against datasets that are inferred under CDM assumptions.