Speaker
Description
Detection of gravitational waves (GWs) has opened new roads in exploring and analyzing astrophysical data. Not only can we learn more about gravitational waves themselves, but this also allows us to perform multi-messenger astronomy, detecting both the GW and electromagnetic (EM) signals. GW170817 demonstrated the power of multi-messenger detections. It confirmed that neutron star mergers produce short gamma-ray bursts, allowed for the observation of the formation and evolution of kilonovae, and enabled the study of the properties of the merger remnant. Additionally, these sources can prove useful in measuring the expansion rate of the universe as well as serving as a test for General Relativity. As of now, only well-modeled GW signals have been detected, but with the increase in sensitivity of the LIGO-Virgo-KAGRA detectors, there is a growing interest in the detection and parameter estimation (PE) of unmodeled signals. Performing multi-messenger detections of unmodeled signals could be even more informative than of modeled signals, revealing mechanisms behind processes that are even less well understood. In this work, we take a step toward addressing this by developing a likelihood-free inference model for sky localization that is designed to generalize beyond its training distribution. We use AMPLFI, a fast, real-time Bayesian parameter estimation ML pipeline. AMPLFI is a PE algorithm based on likelihood-free inference using normalizing flows. We train AMPLFI not on CBC waveforms but on multi–sine Gaussian (multi-SG) signals. These waveforms provide a flexible and general basis for approximating a wide range of transient signals, including those that GWAK may detect. This pipeline is designed to perform quick PE, on the order of a couple of seconds, so that EM follow-up of the signal can be performed.