A neural network built by University of Sydney alumni restored the James Webb Space Telescope’s interferometer to crystal-clear performance, tackling a detector quirk that had been softening fine detail.
📌 Key Takeaways
- Researchers built an AI algorithm that corrects detector distortions in NIRISS/AMI.
- The AI fix improves close-in imaging for exoplanets, jets, and Io’s volcanic surface.
- AMIGO acts in software, no risky telescope servicing or new hardware required.
- The work comes from former University of Sydney PhD students on Tuthill’s team.
- Scientists shared early results on an exoplanet, a cool companion, and WR 137.
What AMIGO Is, And Why It Is AI Not Just “Sharpening”
The team designed AMIGO as a neural network that detects and corrects pixel-level distortions in data from JWST’s Aperture Masking Interferometer.
By learning the detector’s quirks, the AI recovers diffraction-limited structure that standard pipelines left soft, especially at tiny angular separations where faint companions hide.
“Instead of sending astronauts to bolt on new parts, they managed to fix things with code.” — Peter Tuthill, University of Sydney
What Was Blurry, And Why It Mattered For AMI Science
Investigators traced the blur to electronic distortions in JWST’s infrared detector, which skewed the phases and amplitudes that AMI needs for high-contrast imaging.
AMIGO models those non-linear behaviors, then corrects them before interferometric extraction, tightening kernel phases and pushing inner working angles back to design targets.
What Scientists Can See Now That The AI Fix Is In
The team demonstrated sharper views of a dim exoplanet and a very cool, low-mass companion about 133 light-years away.
They also recovered structure in a black hole jet, Io’s volcanic skin, and dust flows from WR 137, all of which demand precise calibration at high contrast.
“This work brings JWST’s vision into even sharper focus.” — Louis Desdoigts, Leiden University
Why A Software Fix Beats A Spacewalk
Hubble needed corrective optics and a shuttle mission. Webb sits near L2, far beyond crewed reach.
With AMIGO, the upgrade happens in the pipeline, not the payload, which lowers mission risk and gives observers a path to adopt improvements without touching hardware.
What’s Next?
Expect updated calibration recipes, broader target lists, and community tests across more filters and epochs. If AMIGO maintains contrast gains, proposals that chase close-in worlds and compact jets will carry more confidence in achievable sensitivity.
Conclusion
JWST did not need a spacewalk to sharpen its interferometer. It needed a precise model of a messy detector and a neural net to unwind it. AMIGO delivers both, and science already looks sharper for it.
The lesson is simple. Calibrate the sensor, not just the sky, and your telescope becomes better overnight.
📈 Latest AI News
5th November 2025
For the recent AI News, visit our site.
If you liked this article, be sure to follow us on X/Twitter and also LinkedIn for more exclusive content.
