Deep learning detection and classification of gravitational waves from neutron star-black hole mergers

• DOI: 10.1016/j.physletb.2023.137850
• Publication Date: 2023-03-16T16:41:19Z
• Authors: Richard Qiu; Plamen G. Krastev; Kiranjyot Gill; Edo Berger
• Keywords: Black hole (networking); Star (game theory)

The Laser Interferometer Gravitational-Wave Observatory (LIGO) and Virgo Collaborations have now detected all three classes of compact binary mergers: black hole (BBH), neutron star (BNS), star-black (NSBH). For coalescences involving stars, the simultaneous observation gravitational electromagnetic radiation produced by an event, has broader potential to enhance our understanding these events, also probe equation state (EOS) dense matter. However, follow-up wave (GW) events requires rapid real-time detection classification GW signals, conventional approaches are computationally prohibitive for anticipated rate next-generation detectors. In this work, we present first deep learning based results signals from NSBH mergers in real LIGO data. We show time that a neural network can successfully distinguish separate them detector noise. Specifically, train convolutional (CNN) on ∼500,000 data samples noise with injected BBH, BNS, high sensitivity accuracy. Most importantly, recover two confirmed to-date (GW191219 GW200115) BNS (GW170817 GW190425), together ∼90% BBH candidate third Gravitational Wave Transient Catalog, GWTC-3. These important step towards low-latency detection, enabling multi-messenger astronomy.