Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Generating gravitational waveform libraries of exotic compact binaries with deep learning

Freitas, Felipe F. ; Herdeiro, Carlos A.R. ; Morais, António P. LU ; Onofre, António ; Pasechnik, Roman LU ; Radu, Eugen ; Sanchis-Gual, Nicolas and Santos, Rui (2024) In Physical Review D 109(12).
Abstract

Current gravitational wave (GW) detections rely on the existence of libraries of theoretical waveforms. Consequently, finding new physics with GWs requires libraries of nonstandard models, which are computationally demanding. We discuss how deep learning frameworks can be used to generate new waveforms "learned"from a simulation dataset obtained, say, from numerical relativity simulations. Concretely, we use the WaveGAN architecture of a generative adversarial network (GAN). As a proof of concept we provide this neural network (NN) with a sample of (>500) waveforms from the collisions of exotic compact objects (Proca stars), obtained from numerical relativity simulations. Dividing the sample into a training and a validation set, we... (More)

Current gravitational wave (GW) detections rely on the existence of libraries of theoretical waveforms. Consequently, finding new physics with GWs requires libraries of nonstandard models, which are computationally demanding. We discuss how deep learning frameworks can be used to generate new waveforms "learned"from a simulation dataset obtained, say, from numerical relativity simulations. Concretely, we use the WaveGAN architecture of a generative adversarial network (GAN). As a proof of concept we provide this neural network (NN) with a sample of (>500) waveforms from the collisions of exotic compact objects (Proca stars), obtained from numerical relativity simulations. Dividing the sample into a training and a validation set, we show that after a sufficiently large number of training epochs the NN can produce from 12% to 25% of the synthetic waveforms with an overlapping match of at least 95% with the ones from the validation set. We also demonstrate that a NN can be used to predict the overlapping match score, with 90% accuracy, of new synthetic samples. These are encouraging results for using GANs for data augmentation and interpolation in the context of GWs, to cover the full parameter space of, say, exotic compact binaries, without the need for intensive numerical relativity simulations.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
in
Physical Review D
volume
109
issue
12
article number
124059
publisher
American Physical Society
external identifiers
  • scopus:85196951560
ISSN
2470-0010
DOI
10.1103/PhysRevD.109.124059
language
English
LU publication?
yes
id
b6cbc437-c5f9-4f3a-8fa1-db49374e2456
date added to LUP
2024-10-03 15:45:39
date last changed
2024-10-03 15:46:30
@article{b6cbc437-c5f9-4f3a-8fa1-db49374e2456,
  abstract     = {{<p>Current gravitational wave (GW) detections rely on the existence of libraries of theoretical waveforms. Consequently, finding new physics with GWs requires libraries of nonstandard models, which are computationally demanding. We discuss how deep learning frameworks can be used to generate new waveforms "learned"from a simulation dataset obtained, say, from numerical relativity simulations. Concretely, we use the WaveGAN architecture of a generative adversarial network (GAN). As a proof of concept we provide this neural network (NN) with a sample of (&gt;500) waveforms from the collisions of exotic compact objects (Proca stars), obtained from numerical relativity simulations. Dividing the sample into a training and a validation set, we show that after a sufficiently large number of training epochs the NN can produce from 12% to 25% of the synthetic waveforms with an overlapping match of at least 95% with the ones from the validation set. We also demonstrate that a NN can be used to predict the overlapping match score, with 90% accuracy, of new synthetic samples. These are encouraging results for using GANs for data augmentation and interpolation in the context of GWs, to cover the full parameter space of, say, exotic compact binaries, without the need for intensive numerical relativity simulations.</p>}},
  author       = {{Freitas, Felipe F. and Herdeiro, Carlos A.R. and Morais, António P. and Onofre, António and Pasechnik, Roman and Radu, Eugen and Sanchis-Gual, Nicolas and Santos, Rui}},
  issn         = {{2470-0010}},
  language     = {{eng}},
  number       = {{12}},
  publisher    = {{American Physical Society}},
  series       = {{Physical Review D}},
  title        = {{Generating gravitational waveform libraries of exotic compact binaries with deep learning}},
  url          = {{http://dx.doi.org/10.1103/PhysRevD.109.124059}},
  doi          = {{10.1103/PhysRevD.109.124059}},
  volume       = {{109}},
  year         = {{2024}},
}