New machine-learning technique gives supermassive black hole at Messier 87 a stunning makeover

The iconic image of the supermassive black hole at the center of Messier 87, a giant elliptical galaxy located 55 million light-years from Earth, has got its first official makeover, thanks to a new machine-learning technique developed by researchers, including an astronomer with NSF's NOIRLab.

The new approach is called PRIMO, which stands for principal-component interferometric modeling, and it relies on a branch of ML known as dictionary learning that teaches computers certain rules by exposing them to thousands of examples. The latest image provides a more comprehensive view of the black hole's dark central region and a narrower outer ring. To accomplish this, a group of scientists utilized the original 2017 data collected by the Event Horizon Telescope (EHT) consortium and generated a new image that showcases the EHT's full resolution, marking the first time this has been achieved.

"With our new machine-learning technique, PRIMO, we were able to achieve the maximum resolution of the current array. Since we cannot study black holes up close, the detail in an image plays a critical role in our ability to understand its behaviour. The width of the ring in the image is now smaller by about a factor of two, which will be a powerful constraint for our theoretical models and tests of gravity, " says lead author Lia Medeiros of the Institute for Advanced Study.

By utilizing the new ML technique on the EHT image of Messier 87, computers analyzed more than 30,000 highly detailed simulated images of gas flowing towards a black hole to identify consistent patterns. These patterns were then combined to create a precise representation of the EHT observations, while also providing an accurate estimation of the missing structures in the image.

"If a picture is worth a thousand words, the data underlying that image have many more stories to tell. PRIMO will continue to be a critical tool in extracting such insights," said Medeiros.

Their findings are published in The Astrophysical Journal Letters.