Scientists led by Northwestern University's Rodrigo Ferrer-Chavez have developed a framework called Physics-based Learning for Imaging Planets (FLIP) that removes starlight from telescope images by combining a physics-based telescope model with optimization techniques used in machine learning. Directly imaging exoplanets is difficult because host stars' glare can overwhelm their faint light, and tiny changes in telescope optics can leave residual glare that hides planets. FLIP models how light passes through the telescope, adjusts for optical imperfections, and simulates the glare to be subtracted. Applied to observations from the James Webb Space Telescope's Near-Infrared Camera (NIRCam), FLIP produced the first image of the known giant planet 51 Eridani b at an infrared wavelength of 4.3 micrometers, where conventional processing had failed. Tests showed sensitivity improvements of two to 15 times over conventional techniques, according to the Nature paper, with the approach improving sensitivity to faint exoplanets by up to 15 times. The paper says this is the first time, to the team's knowledge, that such an approach has demonstrated substantially improved sensitivity on real, on-sky coronagraphic imaging, and that the framework can be extended to other current and future ground-based and space-based observatories. Further work includes reducing computational demands and developing ways to handle the much larger data volumes produced by ground-based observations.