Accelerating Bayesian Sampling for Massive Black Hole Binaries with Prior Constraints from Conditional Variational Autoencoder

Abstract

A Conditional Variational Autoencoder (CVAE) model is employed for parameter inference on gravitational waves (GW) signals of massive black hole binaries, considering joint observations with a network of three space-based GW detectors. Our experiments show that the trained CVAE model can estimate the posterior distribution of source parameters in approximately one second, while the standard Bayesian sampling method, utilizing parallel computation across 16 CPU cores, takes an average of 20 hours for a GW signal instance. However, the sampling distributions from CVAE exhibit lighter tails, appearing broader when compared to the standard Bayesian sampling results. By using CVAE results to constrain the prior range for Bayesian sampling, the sampling time is reduced by a factor of $\sim$6 while maintaining the similar precision of the Bayesian results.

9 pages, 4 figures

References in corpus (10)

• Concepts and status of Chinese space gravitational wave detection projects
• Accelerated gravitational-wave parameter estimation with reduced order modeling
• Neural Importance Sampling for Rapid and Reliable Gravitational-Wave Inference
• Frequency response of space-based interferometric gravitational-wave detectors
• On networks of space-based gravitational-wave detectors
• Identification of Gravitational-waves from Extreme Mass Ratio Inspirals
• A hierarchical search for gravitational waves from supermassive black hole binary mergers
• Constraining Screened Modified Gravity by Space-borne Gravitational-wave Detectors
• Advancing Space-Based Gravitational Wave Astronomy: Rapid Parameter Estimation via Normalizing Flows
• Parameter estimation of microlensed gravitational waves with Conditional Variational Autoencoders

Cited by in corpus (4)

• Fast and accurate parameter estimation of high-redshift sources with the Einstein Telescope
• Accelerating LISA inference with Gaussian processes
• Conditional variational autoencoders for cosmological model discrimination and anomaly detection in cosmic microwave background power spectra
• Simulation-calibrated Bayesian inference for progenitor properties of the microquasar SS 433