Extraction of binary black hole gravitational wave signals from detector data using deep learning

Accurate extractions of gravitational wave signal waveforms are essential to validate a detection and to probe the astrophysics behind the sources producing the gravitational waves. This however, could be difficult in realistic scenarios where the signals detected by the gravitational wave detectors could be contanimnated with non-stationary and non-Gaussian noise. In this paper, we demonstrate for the first time that a deep learning architecture, consisting of Convolutional Neural Network and bi-directional Long Short-Term Memory components can be used to extract all ten detected binary black hole gravitational wave waveforms from the detector data of LIGO-Virgo's first and second science runs with a high accuracy of 0.97 overlap compared to published waveforms.