Astronomers used AI to find 1,400 ‘anomalous objects’ from Hubble archives
The Verge
Astronomers used an AI model called AnomalyMatch to discover nearly 1,400 anomalous objects in the Hubble Space Telescope archives.Summary
European Space Agency (ESA) astronomers David O’Ryan and Pablo Gómez trained an AI model, AnomalyMatch, to systematically search the Hubble Legacy Archive for astrophysical anomalies. The AI scanned nearly 100 million image cutouts in just two and a half days, a task that would overwhelm human teams. The search yielded nearly 1,400 "anomalous objects," including merging galaxies, gravitational lenses, jellyfish galaxies, and several dozen objects that defied classification. The researchers noted this demonstrates AI's utility in maximizing scientific output from large datasets like Hubble's 35-year archive.
(Source：The Verge)