a National Center for Supercomputing Applications, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA
b Department of Physics, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA
c Department of Astronomy, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA
d Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA
DOI: 10.1016/j.physletb.2020.135628
arXiv: arXiv:2004.09524
Published: 14 July 2020
The spin distribution of binary black hole mergers contains key information concerning the formation channels of these objects, and the astrophysical environments where they form, evolve and coalesce. To quantify the suitability of deep learning to characterize the signal manifold of quasi-circular, spinning, non-precessing binary black hole mergers, we introduce a modified version of WaveNet trained with a novel optimization scheme that incorporates general relativistic constraints of the spin properties of astrophysical black holes. The neural network model is trained, validated and tested with 1.5 million \(\ell=|m|=2\) waveforms generated within the regime of validity of NRHybSur3dq8, i.e., mass-ratios \(q\leq8\) and individual black hole spins | \({s}^z_{\{1,\,2\}} | \leq 0.8\). We have deployed a distributed training algorithm at the IBM Power9 Hardware-Accelerated Learning cluster at the National Center for Supercomputing Applications to reduce the training stage from 1 month, using a single V100 NVIDIA GPU, to 12.4 hours using 64 V100 NVIDIA GPUs. Using this neural network model, we quantify how accurately we can infer the astrophysical parameters of black hole mergers in the absence of noise. We do this by computing the overlap between waveforms in the testing data set and the corresponding signals whose mass-ratio and individual spins are predicted by our neural network. We find that the combination of our neural network model and a physics-inspired optimization algorithm enable an accurate reconstruction of the mass-ratio and individual spins of binary black hole mergers across the parameter space under consideration. This is a significant step towards an informed utilization of physics-inspired deep learning models to reconstruct the spin distribution of binary black hole mergers in realistic detection scenarios.
Please refer to our paper for a full discussion of Methods and Results.
Here we provide the interactive version of the figures presented in the paper. You can zoom, hover over plots, and interact with the legend.
Panel 1 summarizes the neural networks performance on the entire test dataset. These subplots correspond to Fig. 8-10 in the paper. Move the slider at the bottom of the panel to access different slices of mass ratio.
Panel 2 corresponds to Fig. 11 in the paper. Here we provide more random samples and user can zoom around t = 0 M to analyze the overlap around merger.
Panel 3 is Fig. A.12 from Appendix A. Hover over points to compare groundtruth and predicted parameters.
Please cite this webpage as
@Misc{KHAN2020135628_interactive_results,
author = "Asad Khan",
title = "Physics-inspired Deep Learning for Gravitational Wave Astrophysics",
year = "2020",
url = "https://khanx169.github.io/smr_bbm_v2/interactive_results.html"
} And cite the paper as
@article{KHAN2020135628,
title = "Physics-inspired deep learning to characterize the signal manifold of quasi-circular, spinning, non-precessing binary black hole mergers",
journal = "Physics Letters B",
pages = "135628",
year = "2020",
issn = "0370-2693",
doi = "https://doi.org/10.1016/j.physletb.2020.135628",
url = "http://www.sciencedirect.com/science/article/pii/S0370269320304317",
author = "Asad Khan and E.A. Huerta and Arnav Das",
}