Rapid Localization of Gravitational Wave Sources from Compact Binary Coalescences Using Deep Learning

Abstract The mergers of neutron star–neutron star and neutron star–black hole binaries (NSBHs) are the most promising gravitational wave (GW) events with electromagnetic (EM) counterparts. The rapid detection, localization, and simultaneous multimessenger follow-up of these sources are of primary importance in the upcoming science runs of the LIGO-Virgo-KAGRA Collaboration. While prompt EM counterparts during binary mergers can last less than 2 s, the timescales of existing localization methods that use Bayesian techniques, vary from seconds to days. In this paper, we propose the first deep learning–based approach for rapid and accurate sky localization of all types of binary coalescences, including neutron star–neutron star and NSBHs for the first time. Specifically, we train and test a normalizing flow model on matched-filtering output from GW searches to obtain sky direction posteriors in around 1 s using a single P100 GPU, which is several orders of magnitude faster than full Bayesian techniques.