Lightning-Fast Gravitational Wave Parameter Inference through Neural Amortization

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.

arXiv commentV1: First version; V2: Updated references; V3: Update references and camera-ready version; V4: Correct figure labels; V5: Updated references

References in corpus (18)

• Advanced Virgo: a 2nd generation interferometric gravitational wave detector
• GWTC-1: A Gravitational-Wave Transient Catalog of Compact Binary Mergers Observed by LIGO and Virgo during the First and Second Observing Runs
• The frontier of simulation-based inference
• Properties of the Binary Black Hole Merger GW150914
• Robust parameter estimation for compact binaries with ground-based gravitational-wave observations using the LALInference software library
• A simple model of complete precessing black-hole-binary gravitational waveforms
• A guide to LIGO-Virgo detector noise and extraction of transient gravitational-wave signals
• Deep Learning for Real-time Gravitational Wave Detection and Parameter Estimation: Results with Advanced LIGO Data
• PyCBC Inference: A Python-based parameter estimation toolkit for compact binary coalescence signals
• Deep Neural Networks to Enable Real-time Multimessenger Astrophysics
• Matching matched filtering with deep networks in gravitational-wave astronomy
• Bayesian parameter estimation using conditional variational autoencoders for gravitational-wave astronomy
• Gravitational-wave parameter estimation with autoregressive neural network flows
• Convolutional neural networks: a magic bullet for gravitational-wave detection?
• Learning Bayesian posteriors with neural networks for gravitational-wave inference
• Approximating Likelihood Ratios with Calibrated Discriminative Classifiers
• Mining for Dark Matter Substructure: Inferring subhalo population properties from strong lenses with machine learning
• Some Optimizations on Detecting Gravitational Wave Using Convolutional Neural Network

Cited by in corpus (14)

• Fast and Credible Likelihood-Free Cosmology with Truncated Marginal Neural Ratio Estimation
• Real-time gravitational-wave inference for binary neutron stars using machine learning
• Learning to Simulate High Energy Particle Collisions from Unlabeled Data
• LtU-ILI: An All-in-One Framework for Implicit Inference in Astrophysics and Cosmology
• Generalized Approach to Matched Filtering using Neural Networks
• Fast and robust Bayesian Inference using Gaussian Processes with GPry
• Swift sky localization of gravitational waves using deep learning seeded importance sampling
• Neural Simulation-Based Inference of the Neutron Star Equation of State directly from Telescope Spectra
• Fast Likelihood-free Reconstruction of Gravitational Wave Backgrounds
• Simulation-based Inference for Gravitational-waves from Intermediate-Mass Binary Black Holes in Real Noise
• Arbitrary Marginal Neural Ratio Estimation for Simulation-based Inference
• Rapid Parameter Estimation for Merging Massive Black Hole Binaries Using Continuous Normalizing Flows
• Group equivariant neural posterior estimation
• Reconsidering the consistent use of precessing, higher order multipole models for gravitational wave analyses