Challenges for Quasar Science with LSST Gordon Richards, Drexel University I will discuss some of the challenges that we face for doing quasar science with LSST, particularly with regard to optimal use of machine learning tools. To find as many quasars as possible (and not to miss rare types) we will have to combine color, variability, astrometric information, and multi-wavelength data and also identify the algorithm(s) that achieve the best completeness and efficiency. To take full advantage of variability data, we must understand how to best make use of light curves in multiple bandpasses and how to parameterize (or not) their variability information. Estimating redshifts presents similar problems to that of quasar discovery: how to best combine all the information at our disposal and what the best photo-z algorithms are. This is particularly important in the regime of low-luminosity AGNs which have properties that are intermediate between galaxies and bona-fide quasars. Lastly, we need to identify the best algorithms for performing clustering and luminosity function analysis that take full advantage of a probabilistic sample, given that LSST lacks a spectroscopic component.