Artificial intelligence lights up black hole fusion

A simulation using an artificial intelligence algorithm succeeds in predicting the characteristics of the fusion of two black holes.

Nearly five years after the discovery of the first gravitational wave in September 2015, a team from the Center for Theoretical Astrophysics of the California Institute of Technology (CalTech, United States) has just published an article which reveals its details , collisions of black holes. Published in Physical Review Letters of January 11, this work presents the most precise simulation to date to describe the fusion of these compact stars.

Machine learning
Thus these researchers laid bare the most cataclysmic event that can occur in the Cosmos: the fusion of two black holes, two extremely compact stars, at the origin of the emission of a gravitational wave. Theoretically predicted by Einstein in 1916, it took physicists a century to invent complex and extremely sensitive detectors such as interferometers capable of detecting the tiny vibrations of space-time that are gravitational waves. For this new study, the researchers used an artificial intelligence algorithm based on machine learning. The learning was done from 900 fusion simulations of two black holes from the Simulating eXtreme Spacetimes (SXS) program calculated on the Wheeler supercomputer from Caltech.

The hunt for new physics

The team manages to predict not only the characteristics of the final black hole, such as its mass and speed of rotation, but also the shape of the gravitational wave that this collision should produce. The use of AI is justified because during the last seconds before the fusion, when the two black holes describe spirals more and more close together, the exact calculation of the shape of the emitted wave is difficult to perform with the traditional methods. The precision of their simulation will be a precious tool for the future generation of detectors: these will be able to surprise small variations in the shape of the wave and possibly track down the tiny effects of new physics.