Conditional noise deep learning for parameter estimation of gravitational wave events

We construct a Bayesian inference deep learning machine for parameter estimation of gravitational wave events binaries black hole coalescence. The structure our adopts the conditional variational autoencoder scheme by conditioning on both strains and variations amplitude spectral density (ASD) detector noise. show that is capable yielding posteriors compatible with ones from nested sampling method better than one without ASD. Our result implies process can be accelerated significantly even large ASD drifting/variation. also apply to LIGO/Virgo O3 events, traditional signal-to-noise ratios higher typical threshold value. use GPU device, NVIDIA RTX3090, train machines, which takes about 12 hours. After training, it less second generate posterior event far faster conventional method.