Actual-time inference for binary neutron star mergers using device learning Evaluation of gravitational waves from merging binary neutron stars was accelerated applying device Mastering, enabling total very low-latency parameter estimation and enhancing the likely for multi-messenger observations.