AnomalyMatch Identifies 1,300+ Hubble Astrophysical Anomalies

Researchers David O'Ryan and Pablo Gómez applied a neural network called AnomalyMatch to 99.6 million Hubble image cutouts and, in two-and-a-half days, identified more than 1,300 anomalous objects, over 800 previously undocumented. Published in Astronomy & Astrophysics and highlighted by NASA on January 27, 2026, discoveries include gravitational lenses, galactic mergers, ring galaxies, and dozens of unclassifiable systems. The work shows AI can rapidly mine archival datasets and prioritize rare phenomena for follow-up.