Finding the brightest QSOs with QUBRICS

The QUBRICS (QUasars as BRIght beacons for Cosmology in the Southern hemisphere) survey is designed to assemble a sample of the brightest quasars with 𝑧>∼ 2.5, observable with facilities in the Southern Hemisphere. The challenge of finding high-redshift QSOs is two-fold: firstly, discriminating QSOs from stars, galaxies and other sources, and, subsequently, estimating a redshift in order to discard low redshift candidates. The strategy devised to address this issue involves the creation of an efficient database, used to train both classification and regression models. QUBRICS takes advantage of Machine Learning methodologies tailored for the extraction of the rare needles-QSOs from the haystack of optical and IR wide-field surveys, optimizing different approaches to maximize completeness or success rates according to specific scientific goals. This talk outlines the lessons learned in the field of machine learning, which hold potential for application across diverse domains beyond astrophysics Furthermore, it presents new results about the evolution of the QSO/AGN luminosity function (in conjunction with recent JWST observations) and the AGN contribution to the cosmic UV background and HI and HeII reionization. Another goal of the QUBRICS project is to take advantage of the newly identified QSOs by using them as probes for the intervening medium. A Golden Sample of seven quasars is presented that makes it possible to carry out the Sandage test of the cosmological redshift drift using the ANDES spectrograph at the 39m ELT with a significant reduction of observing time with respect to previous estimates. Pilot observations have already been initiated with the ESPRESSO spectrograph on the two brightest QSOs that will improve current bounds by an order of magnitude and provide a complete end-to-end proof of concept for the ANDES/ELT experiment as well as a (modest) zeroth epoch.