Scientists have improved the first ever image of a black hole

The famous first image black hole supermassivemade by the EHT collaboration in 2019 in the heart of the galaxy Messier 87was improved by machine learning program.

“Fluffy orange donut” in the first image black hole ever made was reduced to a thinner “thin gold ring” by a program trained on models black holes on a supercomputer.

Better understanding of the phenomenon

Rethinking this image huge black hole at the heart of the galaxy Messier 87 (M87) could help better understand its characteristics and could be expanded to a black hole at the center of our own Milky Way galaxy.

Historical image of hole supermassive black M87 was taken by the Event Horizon Telescope (EHT) and made public in 2019. The EHT collected data to create an image over several days in 2017. The EHT is a network of seven telescopes around the world that creates the Earth. oversized telescope.

A team of researchers, including EHT collaborator and PhD researcher in astrophysics, Leah Medeiros, used a new machine learning technique called Principal Component Interferometric Modeling or “PRIMO” to “fill in the gaps” in the M87 yfr image.

Since we cannot study black holes “Up close, the details of the image play a critical role in our ability to understand its behavior,” Medeiros said in a statement. “The width of the ring in the image is now half as large, which will be a powerful factor. limitation for our theoretical models and gravity tests.”

When the image huge black hole in M87, which is 55 million light-years from Earth and has a mass equivalent to 6.5 billion suns, scientists were surprised at how well it matched Albert Einstein’s 1915 general theory of relativity predictions.

Image Refinement

This is an exquisite image of PRIMO black hole M87 gives scientists the ability to better match observations black hole actual with theoretical predictions.

“PRIMO is a novel approach to the complex task of generating images from EHT observations,” said Tod Lauer, EHT Scientist and NOIRLab researcher. “This provides a way to compensate for the missing information about the observed object, which is necessary to create an image that could be seen with one giant Earth-sized radio telescope.”

The Princeton Institute for Advanced Study explained that PRIMO works using vocabulary learning, a branch of machine learning that allows computers to generate rules from large sets of training materials. So, for example, if such a program is given several images of a banana, it can learn to determine whether an image of an unknown object is a banana.

Teach PRIMO to do the same with black holes, the team provided 30,000 high-fidelity simulated images of these space titans as they fed on ambient gas, a process called “accretion.” The images cover a wide range of theoretical predictions about how black holes they accumulate matter, allowing PRIMO to look for patterns.

Once identified, these patterns were ordered according to the frequency with which they were included in the simulation. This could then be incorporated into EHT images to create a high fidelity image. black hole M87 and discover structures that might have been missed by an array of telescopes.

“We are using physics to fill in missing data areas in a way that has never been used before using machine learning,” Medeiros explained. “This could have important implications for interferometry, which plays a role in fields ranging from exoplanets to medicine.” (Europe Press)

We recommend you METADATA, an RPP technology podcast. News, analytics, reviews, recommendations and everything you need to know about the world of technology.