Deep learning-based gravitational wave detection

A system for detection of gravitational waves, comprising:

• multiple GW detectors that are positioned about the Earth, each of the second plurality of GW detectors comprising components that detect gravitational wave signals, wherein the multiple GW detectors are configured to receive gravitational wave signals and transmit waveform data representing the received gravitational wave signals;
• one or more processors and a memory storing instructions that, when executed by the one or more processors, cause the system to perform the operations of:
• combining time-domain inputs from the multiple GW detectors to improve the accuracy of signal identification by aligning and synchronizing the waveform data transmitted from each of the multiple GW detectors;
• training a deep learning classifier using a first set of waveform data from a first plurality of gravitational wave (GW) detectors;
• applying the trained deep learning classifier to a second set of waveform data comprising the gravitation wave data transmitted from the multiple GW detectors to identify candidate GW signals;
• generating a signal-to-noise ratio (SNR) ranking statistic from the output of the trained deep learning classifier;
• determining SNR rankings of the candidate GW signals using the generated SNR ranking statistic;
• evaluating the compatibility between candidate GW signals from different detectors by comparing arrival times and parameters of the candidate GW signals;
• estimating a plurality of detection significance scores associated with the candidate GW signals by analyzing clusters of identified candidate GW signals that pass or fail the compatibility evaluation;
• and
• outputting at least a subset of the identified GW signals based on their SNR rankings and detection significance scores.