### Astronomers Employ Neural Network to Unveil Secrets of Milky Way's Black Hole

A team of astronomers spearheaded by Michael Janssen from Radboud University in the Netherlands has made groundbreaking advancements in black hole research by leveraging the power of a neural network trained on millions of synthetic data sets. Using data from the Event Horizon Telescope (EHT), this innovative approach has provided new insights into the black hole at the center of our Milky Way, known as Sagittarius A*.

In recent publications in the journal Astronomy & Astrophysics, the researchers revealed that Sagittarius A* is likely spinning at nearly maximum speed, with its rotation axis pointing directly towards Earth. This finding emerges from an extensive analysis that surpasses previous EHT studies, which relied on a limited number of synthetic data files. By employing a Bayesian neural network capable of quantifying uncertainties, the team enhanced the comparison between EHT data and theoretical models.

Notably, the neural network allowed the researchers to infer that the emission near Sagittarius A* is predominantly produced by extremely hot electrons in the surrounding accretion disk, rather than a jet. Furthermore, the magnetic fields within the accretion disk appear to deviate from traditional theoretical expectations. This has led to exciting challenges against established theories, according to lead researcher Michael Janssen.

Janssen and his colleagues see their AI and machine learning methods as a stepping stone to refine and expand models and simulations. The arrival of additional observational data from projects like the Africa Millimeter Telescope, which is currently under construction, promises to bolster these findings and enable more precise tests of the general theory of relativity concerning supermassive compact objects.

The collaborative effort also explored the black hole at the center of galaxy M87, known as M87*. Their findings suggest that M87* is also spinning rapidly but not as swiftly as Sagittarius A*. Interestingly, M87* rotates in the opposite direction to the incoming gas, which could be a remnant of a past merger with another galaxy.

The significant scale of this research was made possible through a coordinated network of computational resources, including CyVerse for data storage, OSG OS Pool for high-throughput computing, Pegasus for workflow management, and the Max Planck Computing and Data Facility for neural network training. Essential software tools like TensorFlow, Horovod, and CASA were crucial in handling the vast amounts of data and complex computations.

The collective effort and use of advanced technology in this project mark a significant milestone in astrophysics, providing deeper insights into the behavior and characteristics of black holes, and paving the way for future discoveries in the field.