Finding new QSOs in photometric catalogs using machine learning: the QUBRICS survey.

The several photometric surveys currently available produced a huge amount of data stored in their catalogs. Besides the primary scientific goals pursued by the individual experiments, there probably is a lot of information hidden in those databases, still waiting to be extracted.
This is certainly the case for the relatively scarce population of high-z QSOs in the Southern Hemisphere known today.

The QUBRICS survey (QUasars as BRIght beacons for Cosmology in theSouthern hemisphere) aims to identify these sources on currently available photometric catalogs using a machine learning approach, and confirm their QSO nature with dedicated spectroscopic observations. More than 200 new, bright (i<18) and high-z (z>2.5) QSOs have already been identified by QUBRICS.

In this talk we will review the performances of the machine learning techniques and the current status of the survey.


Theme – Machine Learning, Statistics, and Algorithms