AI makes supermassive black hole’s first-ever photo sharper. See details

A team of scientists have used artificial intelligence (AI) to make the first-ever picture of a supermassive black hole appear sharper. The first picture of the black hole, also known as ‘Orange Donut’, was shared in 2019. It appeared slightly blurred around the centre, which has now been made sharper and clearer using a machine learning technique known as PRIMO (Principal-component Interferometric Modeling).

The new AI technique was developed by the Event Horizon Telescope (EHT) members, lead author Lia Medeiros of the Institute for Advanced Study, Feryal Ozel and Dimitrios Psaltis of Georgia Tech, and Tod Lauer of NSF’s NOIRLab.

In a statement, Medeiros said that with new machine learning technique PRIMO, the team was able to achieve the maximum resolution. Since the black holes cannot be studied up-close, the detail of an image plays a critical role in the ability to understand its behaviour. The width of the ring in the picture is now smaller, which will be a powerful constraint for theoretical models and tests of gravity. PRIMO will continue to be a vital tool in extracting many insights.


PRIMO used a machine learning technique called dictionary learning where computers analysed over 30,000 high-fidelity simulated images to find common patterns and recreate the earlier EHT image of the Messier 87 black hole.

A black hole is an area of spacetime where gravity is very strong and nothing can escape its event horizon. This particular black hole image captured in 2019 is in the Messier 87 galaxy.
The new image and technique will help scientists and astrophysicists better understand the black hole and its properties. The PRIMO technique will further be used to enhance original images of other space and celestial objects, including the black hole of the Milky Way galaxy.

FAQs:

Q1:When was the first-ever image of the black hole shared with the world?
The original image of the supermassive black hole was shared in 2019.

Q2:What is the full form of PRIMO?
PRIMO is a machine-learning technique. It stands for Principal-component Interferometric Modeling.