We present a spatio-temporal AI framework that concurrently exploits both the spatial and time-variable features of gravitationally lensed supernovae in optical images to ultimately aid in the discovery of such exotic transients in wide-field surveys. Our spatio-temporal engine is designed using recurrent convolutional layers, while drawing from recent advances in variational inference to quantify approximate Bayesian uncertainties via a confidence score. Using simulated Young Supernova Experiment (YSE) images as a showcase, we find that the use of time-series images yields a substantial gain of nearly 20 per cent in classification accuracy over single-epoch observations, with a preliminary application to mock observations from the Legacy Survey of Space and Time (LSST) yielding around 99 per cent accuracy. Our innovative deep learning machinery adds an extra dimension in the search for gravitationally lensed supernovae from current and future astrophysical transient surveys.
Download