Trabajo de grado - Doctorado
Artificial Intelligence in astronomy: machine learning and deep learning approaches to DESI data
Fecha
2023-07-11Registro en:
instname:Universidad de los Andes
reponame:Repositorio Institucional Séneca
Autor
Suárez Pérez, John Fredy
Institución
Resumen
Artificial Intelligence (AI) has shown promise in advancing fundamental physics knowledge, from particle physics to cosmology. The significant advancements in AI over the last decade have been increasingly applied to solve problems in astronomy, primarily motivated by the large amount of data generated by state-of-the-art facilities.
In this thesis, we explore to what extent AI techniques can be useful in tackling problems in observational cosmology. We focused our efforts on applying data mining, machine learning, and deep learning to handle and analyze the data coming from the ongoing observations of the Dark Energy Spectroscopic Instrument (DESI).
DESI is an advanced spectroscopic experiment that has been operational since 2020 and aims to build the most detailed 3D map of the Universe. DESI is a massive undertaking, and over the course of five years, it will measure approximately 40 million spectra from stars, galaxies, and quasars, generating an enormous amount of data that can benefit from advanced AI techniques for analysis and interpretation.
In this thesis, we successfully achieved using AI techniques in three important aspects for DESI:
1) assessing the quality of the data generated by the experiment, 2) describing the cosmic web pattern on the DESI maps, and 3) predicting the redshift observed by DESI from the features observed in imaging data. These three achievements will help improve our understanding of the Universe's evolution and the nature of dark energy, not only with the data coming from DESI but also from future facilities and experiments.