Learning the Universe in the era of Precision Cosmology

I propose the project LUPCOS (Learning the Universe in the era of Precision COSmology) to be carried out at the Kapteyn Astronomical Institute, University of Groningen, under the mentorship of Prof. Pratika Dayal.

The main scientific goal of LUPCOS is to fully capitalize on the upcoming sky-surveys by developing very accurate and highly efficient analytical and semi-numerical models of the Cosmic dawn and the Epoch of Reionization and, finally, employing the latest Machine Learning techniques to obtain the tightest constraints on different cosmological and astrophysical parameters.

The workflow of the project is a combination of machine learning techniques and seminumerical/analytical modelling of the early epochs of the Universe, thereby bridging the gap between these fields to maximise scientific output.

The objectives will be addressed via three work packages (WPs): First (WP1): improve the state-of-the-art analytical model CosmoReionMC by incorporating more observational data and more detailed Physics related to different complex astrophysical processes during CD and EoR.

Second (WP2): Coupling the latest parameter estimation technique with the state-of-the-art semi-numerical galaxy formation model DELPHI to make it suitable for parameter estimation study.

Third (WP3): Employ the most recent machine learning technique on the state-of-the-art cosmological N-body simulation data to extract the maximum information; this will work as a precursor project on parameter inference techniques for the upcoming sky surveys.

My expertise in analytical modelling of the early Universe and skills in ML techniques, along with the expertise of my proposed supervisor, Prof. P. Dayal, at the Kapteyn, put us in a favourable position for the successful execution of this project.

Furthermore, with the training from this fellowship, I will be competent to lead future research groups of my own, paving the way for ground-breaking scientific results with upcoming surveys.