Free Astronomy Magazine May-June 2026

documented in the scientific litera- ture. Most of the anomalies were galaxies in the process of merging or interacting, taking on unusual shapes or trailing long tails of stars and gas. Many others were gravita- tional lenses, in which the gravity of a foreground galaxy bends space- time and warps the light from a dis- tant background galaxy into a circle or arc. The team also discovered ex- amples of several other rare objects such as galaxies with huge clumps of stars, jellyfish galaxies with gaseous ‘tentacles’, and planet-forming disks seen edge-on, giving them a ham- burger-like or butterfly-like appear- ance. Perhaps most intriguing of all, there were several dozen objects that defied classification altogether. “This is a fantastic use of AI to maximise the scientific output of the Hubble archive,” says study co- author Pablo Gómez. “Finding so many anomalous objects in Hubble data, where you might expect many to have already been found, is a great result. It also shows how useful this tool will be for other large datasets.” search for objects that are outside the norm — or find them by chance. While trained scientists excel at spot- ting cosmic anomalies, there’s simply too much Hubble data for experts to sort through at the necessary level of fine detail by hand. Citizen science projects, which enlist non-scientists to collaborate on tasks such as classifying galaxies, provide another way to chip away at the mountains of data available. While citizen science groups greatly expand the amount of data that can be inspected, they’re still no match for extensive archives like Hubble’s, or for datasets from telescopes that survey the sky like ESA’s Euclid space telescope. Now, this new work by O’Ryan and Gómez takes the search to a whole new level. The team devel- oped what’s called a neural network, an AI tool that uses comput- ers to process data and search for patterns in a way that is inspired by the human brain. Their neural network, which they named Anomaly- Match, is trained to search for and recog- nise rare objects like jellyfish galaxies and gravitational arcs. 41 MAY-JUNE 2026 ASTRO PUBLISHING P revious page: Six previously-undiscovered, weird and fascinating astrophysical objects are displayed in this new image from the NASA/ESA Hubble Space Telescope. These were discovered by researchers from the European Space Agency using a new AI-assisted method. The AI tool allowed them to search nearly 100 million image cutouts and uncover anom- alous objects including gravitational lenses, jellyfish galaxies with gaseous ‘tentacles’, merging and interacting galaxies, galaxies featuring rings and arcs and more. This collection features six galaxies, showing a cross-section of the discoveries with some of the more striking examples: three lenses with arcs distorted by gravity, one galactic merger, one ring galaxy, and one galaxy — not alone in the results — which defied classification. To detect anomalous objects like these six, the re- searchers developed an AI tool capable of searching and recognising patterns in images, and trained it with examples of types of unusual objects that they wanted to find. They then used their algorithm to examine the entire set of data from Hubble’s archive in search of further anomalous objects, over the course of just a couple of days. The result was a ranking of which images contained objects most likely to be anomalous. After inspecting the results from their AI tool, the team confirmed more than 1300 anomalies, of which over 800 were previously unknown. With even larger datasets on the way from missions including ESA’s Euclid, the hope is that AI tools such as this one can help astronomers to make the absolute most of their observations. [ESA/Hubble & NASA, D. O’Ryan, P. Gómez (European Space Agency), M. Zamani (ESA/Hubble)] The team used AnomalyMatch to search through nearly 100 million image cutouts from the Hubble Legacy Archive, marking the first time the archive has been systemat- ically searched for astrophysical anomalies. In just two and a half days, AnomalyMatch completed its search of the archive and returned a list of likely anomalies. As the process of tracking down rare objects still requires an expert eye, O’Ryan and Gómez personally inspected the sources rated by their algorithm as most likely to be anomalous. Of these, more than 1300 were true anomalies, more than 800 of which had never been https://esahubble.org/videos/heic2603a/ T his Space Sparks episode highlights how a team of astronomers have used a new AI-assisted method to search for rare astronomi- cal objects in the Hubble Legacy Archive. [ESA/Hubble & NASA] !