AI and Hubble Science - NASA Science

Suggested Searches

Huge sets of data pulled in from telescopes and observatories around the world, including the Hubble Space Telescope, provide a treasure trove of information for astronomers seeking to unlock the secrets of the cosmos. But while it would take countless hours for individuals to sort through information from years of observations, artificial intelligence (AI) programs can use pattern recognition to swiftly identify key components for astronomical investigations.

Astronomers have used AI and Hubble data to hunt asteroids in between the orbits of Mars and Jupiter. These small asteroids are faint and difficult to detect, but leave distinctive curved, streak-like trails on Hubble’s observations. Astronomers used machine learning algorithms, a form of AI, to identify these streaks in over 30,000 Hubble images. Combined with the efforts of some 11,000 citizen scientist volunteers, the project revealed 1,031 previously undiscovered asteroids, providing valuable information about the formation and evolution of our solar system’s asteroid belt.

Hundreds of galaxies of various colors and sizes are scattered across the view, and about a dozen curved or wavy white streaks appear in random locations.

Curved, streak-like trails left by asteroids on Hubble images are one of the items that artificial intelligence programs can sort through vast amounts of data to help identify.

NASA, ESA, and B. Sunnquist and J. Mack (STScI) Acknowledgment: NASA, ESA, and J. Lotz (STScI) and the HFF Team

Similarly, Hubble observations of thousands of galaxies helped train AI programs to identify galaxy structures and forms – sometimes on a pixel-by-pixel basis. Researchers use such programs on data like the Hubble Legacy Field, a combination of nearly 7,500 separate Hubble exposures, representing 16 years of observations that contains over 265,000 galaxies, to speed through galaxy classifications. They also used Hubble data to test machine learning algorithms that remove flaws and dust clouds from “noisy” Hubble images, clarifying and refining images to reveal obscured details.

Six Hubble images of distorted galaxies are organized in a two-row mosaic. From left to right, the top row of galaxies appears as follows: The left panel has a galaxy that resembles the number nine tilted on its side to the left and has red-orange regions scattered with blue knots. The center square shows an edge-on spiral galaxy appearing like a white thin bar extending from 8 o’clock to 2 o’clock. It has a bright, compact core and a small background spiral galaxy just below the core. The right panel shows two merging galaxies forming a convoluted shape that extends from 8 o’clock to 2 o’clock. The bottom row of galaxies appears as follows: Left square contains a face-on spiral with faint, broad arcs of material to its left and right. The center panel has a hazy white, face-on spiral with a lumpy vertical line to its right that appears to curve around its core. The right panel shows an orange elliptical galaxy with a lumpy blueish galaxy curving around it to the right.

Six previously undiscovered, weird, and fascinating astrophysical objects are displayed in this image from NASA’s Hubble Space Telescope. They include three lenses with arcs distorted by gravity, one galactic merger, one ring galaxy, and one galaxy that defied classification. Researchers found these galaxies using AI to analyze 100 million image cutouts from the Hubble Legacy Archive.

Image: NASA, ESA, David O'Ryan (ESA), Pablo Gómez (ESA), Mahdi Zamani (ESA/Hubble)

One team analyzed nearly 100 million image cutouts from the Hubble Legacy Archive, each measuring just a few dozen pixels (7 to 8 arcseconds) on a side. They identified more than 1,300 objects with an odd appearance in just two and a half days — more than 800 of which were undocumented in scientific literature. (To learn more, see: AI Unlocks Hundreds of Cosmic Anomalies in Hubble Archive)

As observatories become increasingly effective, their data collections also grow to create enormous archives of observations. With the help of AI programs that can sort through vast troves of data in search of identifiable patterns, researchers are poised to make new leaps in scientific discovery.

More Things Hubble

nearly spherical gathering of white stars at center top. Blue gas cloud extends from top left to center bottom

Hubble Science Highlights

Hubble’s most notable scientific discoveries reflect the broad range of research and the breakthroughs it has achieved.

Hubble Space Telescope

Hubble Oral Histories

These oral history interviews document the experiences of individuals who were involved in every aspect of NASA's Hubble Space Telescope mission.

Upper left: "Hubble Focus" opening of the telescope is the "O" Center: bright-white star, Center Bottom: Orange-red planet with cloud bands, Center: small, bright-white star and very small red-orange dwarf star

Hubble E-books

Investigate the mysteries of the universe with Hubble. Learn Hubble's history. E-books dive deeper into Hubble discoveries and more.

Hubble Observatory - Diagram

Observatory Design

Hubble is the first space-based observatory specifically designed for servicing while in orbit. Learn more about its systems and design.

Hubble Timelines

Follow Hubble's journey from an idea of a "Large Space Telescope" to its position as one of the most scientifically productive and beloved telescopes of all time. Also follow its timelines of discoveries and servicing missions.

Keep Exploring

Explore More

Hubble Space Telescope

Since its 1990 launch, the Hubble Space Telescope has changed our fundamental understanding of the universe.

3D Hubble Models

Hearing Hubble

Hubble’s Night Sky Challenge