🌌 AI Model Predicts the Rotation Speed of a Black Hole at the Center of the Milky Way
An international team of scientists has created the first AI model capable of accurately predicting the properties of black holes based on raw data from telescopes.
Astronomers created 962,000 synthetic "images" of black holes based on existing physical models, adding distortions and errors that inevitably happen during real observations. Then, they trained an AI model using this data.
The researchers then loaded the Event Horizon Telescope network data into the model. Using this decentralized observatory, astronomers obtained the first photo of a black hole in the center of the galaxy M87 in 2019, and in 2022, an image of the black hole Sagittarius A* (Sgr A*) in the center of our Milky Way ⤴️ But a lot of data remained unprocessed—and the new model was able to extract valuable information from it.
🔭 It turned out that Sgr A* rotates at almost the maximum speed allowed by theory, and its axis of rotation is directed toward Earth. The simulation also showed that the behavior of the black hole's accretion disk is more complex than the predicted behavior of existing models.
🔭 The M87 black hole rotates in the opposite direction to the spin of its accretion disk. Astronomers believe this may result from a merger with another galaxy many years ago.
"That we are defying the prevailing theory is of course exciting," says the project's lead researcher Michael Janssen. He is confident that AI will help us study black holes even better, especially when we can obtain more high-quality data for AI training with new telescopes.
