The discovery of the first strong gravitational lensed quasar brought forth a powerful observational tool for cosmology and astrophysics. Systems where the background source is a quasar can be used to map the dark matter substructure, determine the mass and spin of black holes, measure the properties of quasar host galaxies, and measure the value of the Hubble constant H0. In addition, the effects of microlensing of the background quasar flux induced by the stars in the lens galaxy can be used to probe the physical properties of quasar accretion disks.
This talk will present a new morphology independent photometric quasar selection technique and the subsequent selection of gravitationally lensed quasars candidates, in contrast to previous studies which have mostly focus on spectroscopically confirmed quasars or radio selection of candidates. The method uses supervised machine learning (specifically Gaussian Mixture Models - GMM) to select quasar candidates on multi-dimensional colour space, mixing optical and infrared data. Then, lensed quasar systems are selected through a combination of cuts in morphological properties and 2D image modelling of the quasar candidates in the GMM generated catalogues. In the end, I will show the new systems discovered in Dark Energy Survey data.