Lightning-Fast Gravitational Wave Parameter Inference through Neural Amortization

TL;DR

This paper introduces a neural amortization method that drastically accelerates gravitational wave parameter inference, reducing computation time from days to minutes while maintaining accuracy.

Contribution

It presents a novel neural simulation-based inference approach that models the likelihood-to-evidence ratio to efficiently estimate gravitational wave parameters.

Findings

Inference time reduced by up to three orders of magnitude.

The neural model accurately estimates credible intervals.

Performance comparable to traditional methods.

Abstract

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.