Lightning-Fast Gravitational Wave Parameter Inference through Neural Amortization. (arXiv:2010.12931v5 [astro-ph.IM] UPDATED)
<a href="http://arxiv.org/find/astro-ph/1/au:+Delaunoy_A/0/1/0/all/0/1">Arnaud Delaunoy</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Wehenkel_A/0/1/0/all/0/1">Antoine Wehenkel</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Hinderer_T/0/1/0/all/0/1">Tanja Hinderer</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Nissanke_S/0/1/0/all/0/1">Samaya Nissanke</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Weniger_C/0/1/0/all/0/1">Christoph Weniger</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Williamson_A/0/1/0/all/0/1">Andrew R. Williamson</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Louppe_G/0/1/0/all/0/1">Gilles Louppe</a>
Gravitational waves from compact binaries measured by the LIGO and Virgo
detectors are routinely analyzed using Markov Chain Monte Carlo sampling
algorithms. Because the evaluation of the likelihood function requires
evaluating millions of waveform models that link between signal shapes and the
source parameters, running Markov chains until convergence is typically
expensive and requires days of computation. In this extended abstract, we
provide a proof of concept that demonstrates how the latest advances in neural
simulation-based inference can speed up the inference time by up to three
orders of magnitude — from days to minutes — without impairing the
performance. Our approach is based on a convolutional neural network modeling
the likelihood-to-evidence ratio and entirely amortizes the computation of the
posterior. We find that our model correctly estimates credible intervals for
the parameters of simulated gravitational waves.
Gravitational waves from compact binaries measured by the LIGO and Virgo
detectors are routinely analyzed using Markov Chain Monte Carlo sampling
algorithms. Because the evaluation of the likelihood function requires
evaluating millions of waveform models that link between signal shapes and the
source parameters, running Markov chains until convergence is typically
expensive and requires days of computation. In this extended abstract, we
provide a proof of concept that demonstrates how the latest advances in neural
simulation-based inference can speed up the inference time by up to three
orders of magnitude — from days to minutes — without impairing the
performance. Our approach is based on a convolutional neural network modeling
the likelihood-to-evidence ratio and entirely amortizes the computation of the
posterior. We find that our model correctly estimates credible intervals for
the parameters of simulated gravitational waves.
http://arxiv.org/icons/sfx.gif
