Orateur
Bhaskar arya
(IIT Kanpur)
Description
We present a machine learning framework that uses neural networks to predict Lyman-$\alpha$ forest spectra from dark matter density fields. Trained on simulations with varying cosmological parameters, our network learns the complex, non-linear transformation from the underlying matter distribution and velocity fields to transmitted flux. We demonstrate that the network accurately reconstructs the Lyman-$\alpha$ absorption features along sightlines within the same simulation box and speculate that this approach can enable fast and accurate generation of Lyman-$\alpha$ spectra, offering a powerful tool for emulating large-scale structure observables and accelerating parameter inference in cosmology.
Auteur
Bhaskar arya
(IIT Kanpur)
