December 16, 2018 Deep-learning technique reveals “invisible” objects in the dark, requiring only one photon per pixel