Inference of weak gravitational lensing signals from sky images
Weak gravitational lensing is a highly valued tool for inferring the structure of the spacetime metric between an observer and a cosmologically distant “wallpaper,” most commonly either the CMB or faint background galaxies. The best-measured quantities are the second derivatives of the projected scalar potential(s), which are manifested as apparent shearing and magnification of the wallpaper. Given a collection of faint-galaxy images, what information can we extract about the shear and magnification that these images have undergone? I will describe a new method of lensing inference that, unlike predecessors, is rigorously correct in the presence of noise and other observational realities, nearly optimal, and computationally feasible at the scale of current/future surveys like the Dark Energy Survey and the Large Synoptic Survey Telescope.