Self-learning neural network cracks iconic black holes

Stephanie Baum
scientific editor

Robert Egan
associate editor

A team of astronomers led by Michael Janssen (Radboud University, The Netherlands) has trained a neural network with millions of synthetic black hole data sets. Based on the network and data from the Event Horizon Telescope, they now predict, among other things, that the black hole at the center of our Milky Way is spinning at near top speed.
The astronomers have published their results and methodology in the journal Astronomy & Astrophysics.
In 2019, the Event Horizon Telescope Collaboration released the first image of a supermassive black hole at the center of the galaxy M87. In 2022, they presented an image of the black hole in our Milky Way, Sagittarius A*. However, the data behind the images still contained a wealth of hard-to-crack information. An international team of researchers trained a neural network to extract as much information as possible from the data.
From a handful to millions
Previous studies by the Event Horizon Telescope Collaboration used only a handful of realistic synthetic data files. This time, the astronomers fed millions of such data files into a so-called Bayesian neural network that can quantify uncertainties. This allowed the researchers to make a much better comparison between the EHT data and the models.
Thanks to the neural network, the researchers now suspect, for example, that the black hole at the center of the Milky Way is spinning almost at top speed. Its rotation axis points to Earth. In addition, the emission near the black hole is mainly caused by extremely hot electrons in the surrounding accretion disk and not by a so-called jet. Also, the magnetic fields in the accretion disk seem to behave differently than the usual theories of such disks.
"That we are defying the prevailing theory is of course exciting," says lead researcher Michael Janssen (Radboud University Nijmegen, the Netherlands). "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. And when the Africa Millimeter Telescope, which is under construction, joins in with data collection, we will get even better information to validate the general theory of relativity for supermassive compact objects with a high precision."
Impressive scaling
"The ability to scale up to millions of synthetic data files is an impressive achievement," says co-researcher Jordy Davelaar (Princeton University, U.S.). "You need storage capacity, a supercomputer, a software pipeline, and a program that distributes the work."
The researchers stress that this scale of work was made possible by a coordinated ecosystem of computational services: CyVerse for data storage, OSG OS Pool for high-throughput computing, Pegasus for workflow management, Germany's Max Planck Computing and Data Facility for the neural network training, and software tools including TensorFlow, Horovod, and CASA.
The researchers did not just make predictions about Sagittarius A*. They also looked at M87*, the black hole at the center of M87. Among other things, they found that this black hole is also spinning fast, but not as fast as Sagittarius A*. Besides that, it is spinning in the opposite direction to the infalling gas. The astronomers suggest that this counter-rotating motion may be the result of a merger with another galaxy.
More information: M. Janssen et al, Deep learning inference with the Event Horizon Telescope I. Calibration improvements and a comprehensive synthetic data library, Astronomy & Astrophysics.
M. Janssen et al, Deep learning inference with the Event Horizon Telescope II. The Zingularity framework for Bayesian artificial neural networks, Astronomy & Astrophysics.
M. Janssen et al, Deep learning inference with the Event Horizon Telescope III. Zingularity results from the 2017 observations and predictions for future array expansions, Astronomy & Astrophysics.
Journal information: Astronomy & Astrophysics
Provided by Netherlands Research School for Astronomy