Speaker
Description
This talk presents a research project centered on the construction and scientific exploitation of the galaxy and quasar catalogs in J-PAS and S-PLUS surveys. Building upon expertise developed within the S-PLUS survey, this project utilizes Machine Learning (ML) techniques - such as Bayesian Neural Networks, FlexCoDE, and foundation models - to enhance classification and photometric redshift estimation of galaxies and quasars by combining narrow-band and broad-band filters with near-infrared data. A central component of the research involves developing and validating new ML tools. We will present an analysis to identify systematics, compute galaxy and quasar bias, allowing us to use the large-scale structure traced by these objects to place constraints on cosmological parameters.