
AI Discovers 1,300 Hidden Anomalies in Hubble Archive
ESA researchers used a neural-network tool called AnomalyMatch to scan nearly 100 million tiny image snippets from the 35-year Hubble archive, identifying over 1,300 anomalies—more than 800 of which were not previously documented—mostly related to galactic mergers, jellyfish galaxies, edge-on planet-forming disks, and gravitational lenses; the work demonstrates AI’s power to boost discoveries from archival space data.






