Kilonova Light Curve Interpolation with Neural Networks

A kilonova is an astronomical transient that occurs after the merger of two neutron stars or a neutron star and a black hole. The kilonova is powered by the decay of radioactive elements on timescales of hours to weeks post-merger, emitting energy in the form of observable light curves. In modeling kilonova light curves, we use radiative transfer simulations to study the fundamental physics in these complex environments. However, these methods are slow and computationally expensive, prompting the use of surrogate modeling techniques such as neural networks. In this poster, we describe a neural network trained on existing kilonova lightcurve simulations. Our neural network emulator can generate millions of new light curves in minutes when trained on a set of 22248 radiative transfer simulations which took weeks to simulate. We also present our network’s successful recovery of kilonova light curves in our test set, motivating parameter inference application as discussed in an associated talk.