Generating higher order modes from binary black hole mergers with machine learning

Tim Grimbergen,Stefano Schmidt,Chinmay Kalaghatgi,Chris van den Broeck

DOI: https://doi.org/10.1103/physrevd.109.104065

IF: 5.407

2024-05-21

Physical Review D

Abstract:We introduce a machine learning model designed to rapidly and accurately predict the time domain gravitational wave emission of nonprecessing binary black hole coalescences, incorporating the effects of higher order modes of the multipole expansion of the waveform. Expanding on our prior work [ Phys. Rev. D 103 , 043020 (2021)], we decompose each mode by amplitude and phase and reduce dimensionality using principal component analysis. An ensemble of artificial neural networks is trained to learn the relationship between orbital parameters and the low-dimensional representation of each mode. Our model is trained with ∼105 signals with mass ratio q∈[1,10] and dimensionless spins χi∈[−0.9,0.9] , generated with the state-of-the-art approximant seobnrv 4 hm , and it is able to generate waveforms up to ∼4×105M long. We find that it achieves a median faithfulness of 10−4 averaged across the parameter space. We show that our model generates a single waveform 2 orders of magnitude faster than the training model, with the speedup increasing when waveforms are generated in batches. This framework is entirely general and can be applied to any other time domain approximant capable of generating waveforms from aligned spin circular binaries, possibly incorporating higher order modes. https://doi.org/10.1103/PhysRevD.109.104065 © 2024 American Physical Society

astronomy & astrophysics,physics, particles & fields