Hunting for electromagnetic counterparts of compact binary mergers using federated learning

Next-generation gravitational-wave observatories (Cosmic Explorer and Einstein Telescope) are expected to bring orders of magnitude more detections of binary neutron star mergers compared to what is achievable today by LIGO and Virgo. Here, we present a framework aimed at streamlining the problem of finding electromagnetic counterparts to gravitational wave events, leveraging AI and federated learning techniques. Our goal is to improve the efficiency in the electromagnetic follow up across a variety of wavelengths and data formats, while preserving data proprietary rights. We conclude by considering methods of ensuring the resilience of this framework across distributed resources.