It is possible that there may be a problem AI will soon be able to unlock the mysteries of these supermassive objects. An international team has trained a network of neural networks with billions of simulations of black holes to enable it to understand fuzzy data collected from these mysterious objects.
The Event Horizon Telescope, which is one of the many methods to investigate a blackhole, is probably the best known. EHT, which is not one instrument, is a group of radiotelescopes located around the world working together as a telescope. Thanks to the EHT, it’s been possible to obtain images of the supermassive black holes M87 and Sagittarius A*. The images aren’t traditional pictures, but rather visualizations of the radio waves that are coming from black holes.
Supercomputers around the globe processed radio signals from the EHT to create the images. In the end, however, much of the data was thrown away because it was hard to interpret. Experts at Morgridge Research Institute, Wisconsin, have trained a new neural net to use this data sea to make discoveries and improve EHT’s readings.
According to a press release from the institute, the artificial intelligence successfully analyzed the once-discarded information and established new parameters of Sagittarius A*, which sits at the center of the Milky Way. A new image was created of the structure of the blackhole, revealing some of its characteristics.
“Researchers now suspect that the black hole at the center of the Milky Way is spinning at almost top speed,” The researchers wrote in a press release. The image shows that the blackhole’s rotation axis is pointing towards Earth. It also gives clues about the origins and characteristics of discs of material which circulate around the hole.
Astronomers had previously estimated that Sagittarius A* rotates at a moderate to fast speed. The actual speed of rotation is critical, because it can be used to determine how the radiation surrounding the black hole will behave and gives clues as to its stability.
“That we are defying the prevailing theory is of course exciting,” Michael Janssen from Radboud University Nijmegen (Netherlands) is the principal researcher, according to a recent press release. “However, I see our AI and machine learning approach primarily as a first step. Next, we will improve and extend the associated models and simulations.”
The original version of this story appeared originally on WIRED en Español This has been translated into Spanish.
