Graph Neural Networks for Interferometer Simulations

Graph neural networks can model the complex optical physics of gravitational-wave interferometers while running substantially faster than conventional simulation tools. This work also introduces a high-fidelity dataset spanning three interferometer topologies.