Researchers have created a novel machine learning method to analyse gravitational waves emitted from neutron star collisions almost instantaneously – even before the merger is fully observed.
Observing binary neutron star mergers is high on the wish list of astronomers. These collisions of exotic, compact stellar remnants emit gravitational waves followed by light, providing unique opportunities to study gravity and matter under extreme conditions. These multi-messenger observations depend crucially, however, on fast analysis of gravitational wave data, a computationally demanding task.
In a study published in the journal Nature, an interdisciplinary team of researchers presents a novel machine learning method that could be instrumental in preparing the field for the next generation of observatories.
The international team of scientists, including UKRI Future Leaders Fellow Stephen Green from the University of Nottingham, developed an algorithm called DINGO-BNS (Deep INference for Gravitational-wave Observations from Binary Neutron Stars). It trains a neural network to fully characterize systems of merging neutron stars in about a second, enabling a fast search for visible light and other electromagnetic signals emitted during the collisions. This compares to about an hour for the fastest traditional methods. Their paper is titled "Real-time inference for binary neutron star mergers using machine learning".
Neutron star mergers emit gravitational waves, visible light (in an explosion known as a kilonova) and other electromagnetic radiation, as shown in this video. "Rapid and accurate analysis of the gravitational-wave data is crucial to localize the source and point telescopes in the right direction as quickly as possible to observe all the accompanying signals," says Maximilian Dax, a Ph.D. student in the Empirical Inference Department at the Max Planck Institute for Intelligent Systems (MPI-IS) and first author of the paper.
Gravitational wave analysis is particularly challenging for binary neutron stars, so for DINGO-BNS, we had to develop various technical innovations. This includes for example a method for event-adaptive data compression.
The real-time method sets a new standard for data analysis of neutron star mergers, giving the broader astronomy community more time to point their telescopes toward the merging neutron stars as soon as the gravitational-wave detectors of the LIGO-Virgo-KAGRA (LVK) collaboration identify them.
"Current rapid analysis algorithms used by the LVK make approximations that sacrifice accuracy. The new study addresses these shortcomings," says Jonathan Gair, a group leader in the Astrophysical and Cosmological Relativity Department at the Max Planck Institute for Gravitational Physics in the Potsdam Science Park. Bernhard Schölkopf, Director of the Empirical Inference Department at MPI-IS, adds: "Our study showcases the effectiveness of combining modern machine learning methods with physical domain knowledge."
Indeed, the machine learning framework fully characterizes the neutron star merger (e.g., its masses, spins, and location) in just one second without making such approximations. This allows, among other things, to quickly determine the sky position 30% more precisely. Because it works so quickly and accurately, the neural network can provide critical information for joint observations, and make the best possible use of the expensive telescope observing time.
DINGO-BNS could one day help to observe electromagnetic signals even before the collision of the two neutron stars. "Such early multi-messenger observations could provide new insights into the merger process and the subsequent kilonova, which are still mysterious," says Alessandra Buonanno, Director of the Astrophysical and Cosmological Relativity Department at the Max Planck Institute for Gravitational Physics.
