AI Assisted in Creating First All-Sky Ultraviolet Map
Anthropic announced its Claude Science AI model assisted in mapping the first complete ultraviolet map of the entire sky. Approximately one-third of the map's area was estimated by the model.
An artificial intelligence model has aided astronomers in generating the first comprehensive ultraviolet map of the entire sky. Anthropic confirmed that its Claude Science AI was instrumental in estimating roughly one-third of the map, covering areas previously lacking UV observations.
The complete all-sky UV map combines data from both far-UV (154 nm) and near-UV (232 nm) wavelengths centered around the Milky Way's core. Earth's ozone layer prevents ground-based UV observation, necessitating the use of space telescopes. While NASA's GALEX mission provided the largest dataset, about a third of the sky remained unobserved.
The research team integrated data from multiple space and satellite missions, including GALEX, NASA's Swift, and Korea's FIMS/SPEAR. For the unobserved regions, Claude Science estimated UV brightness by analyzing its correlation with UV levels in observed areas and data from visible light, infrared, and radio waves.
The resulting map distinguishes between directly measured and AI-estimated pixels, with the latter covering substantial portions of the galactic disk. It also provides uncertainty levels for the estimations, offering the scientific community a new tool for galactic research.