Campus, Varian 355
The field of CMB lensing is somewhere akin to where measurements of the primary CMB itself were 15 years ago; we have detected it's there and measured some scales to moderate significance, but the exciting era of deep precision measurements is just on the horizon. Over the coming decade, CMB lensing (the distortion of the CMB photons by gravitational lensing due to the matter along the CMB photon's paths) will precisely probe the distribution of matter in the universe out to high red shifts. It will also allow us to remove lensing B-mode polarization, which will be crucial to potentially detecting primordial gravitational waves from the Big Bang. However, perhaps surprisingly given the important role CMB lensing will play, how to fully extract all of the lensing information from precision CMB data is an open question, as the lensing reconstruction is a difficult high-dimensional non-linear problem. In this talk, I'll provide a review of CMB lensing science aimed at non-experts. I'll also describe the methods we have been developing to perform this optimal extraction, which are rooted in Bayesian statistics, machine learning, and modern optimization and sampling algorithms.