The Bottleneck Nobody Talks About

The James Webb Space Telescope is the most expensive and sought-after instrument on Earth. But it has a brutal limitation: there are only so many hours in a year, and demand vastly exceeds supply. Every minute observing one galaxy is a minute not spent on another [1].

Now, a team led by Yuduo Guo at Tsinghua University has found a clever workaround—and it works remarkably well.

Enter ASTERIS: The Noise Assassin

Known as ASTERIS, the AI network removes noise from images to reveal features a full magnitude fainter than before. [2] Think of it as digital enhancement that's more sophisticated than sharpening a blurry photo.

For a proof of concept, ASTERIS has already more than doubled the number of distant galaxies detected in a set of images taken by the James Webb Space Telescope. [2] Those aren't new observations—they're existing data, now seen with new clarity.

Here's the practical magic: A real image made from 168 exposures resulted in 21 times the exposure time, yet most of the sources in the ASTERIS-processed image made from just eight exposures were real objects, as revealed by the deeper image. [2]

What Makes This Credible (Not Vaporware)

In AI, the danger is always "hallucinations"—the system inventing galaxies that don't exist. The team tackled this head-on. They "inject" fake signals to see if ASTERIS can pick them up, then train the network on less data than what's actually available to see how it does. [2]

The real test came with verification: The team, led by Yuduo Guo (Tsinghua University, China), tried multiple ways to verify what ASTERIS found in Webb images. [2] The algorithm published in Science, meaning peer reviewers scrutinized the claims.

Why This Matters Beyond Headlines

This is a pattern we're seeing across astronomy: expensive, limited instruments (Webb, Hubble, ground-based observatories) are being paired with AI to extract more signal from existing data. A team of astronomers using a new AI-assisted method searched the Hubble Legacy Archive, sifting through nearly 100 million image cutouts in just two and a half days, uncovering nearly 1,400 anomalous objects, more than 800 of which had never been documented before. [3]

The implications are concrete: if you can effectively multiply the value of each observation hour, you change the calculus of what's scientifically reachable.

What's Next

The team is planning to move forward, applying ASTERIS to additional data from both Webb as well as the 8.2-meter Subaru Telescope in Hawai'i. [2] With the network's design published, independent verification will likely follow—and adoption across other observatories.

This isn't glamorous. It won't grab headlines like finding the most distant galaxy or glimpsing organic molecules in a distant galaxy. But for working astronomers drowning in observational data yet starved for observation time, ASTERIS represents something more valuable: a genuine multiplication of capability.

Sources & References

[1] NASA Science - James Webb Space Telescope: https://science.nasa.gov/mission/webb/

[2] Sky & Telescope - "AI Reveals New Galaxies in James Webb Space Telescope Images" (2026): https://skyandtelescope.org/astronomy-news/ai-reveals-new-galaxies-in-james-webb-space-telescope-images/

[3] ESA/Hubble - Press Releases "AI finds hundreds of cosmic anomalies in Hubble Legacy Archive": https://esahubble.org/news/