Classifying Gamma-Ray Burst X-Ray Afterglows with a Variational Autoencoder

Abstract

Today, over 1500 gamma-ray burst (GRB) afterglows have been observed by The Neil Gehrels Swift Observatory, with light curves displaying different morphologies in the succession of decay regimes with time. We explore prospects for acquiring physical inference from ML models by investigating the presence of different classes in the morphology of GRB afterglow X-ray light curves. We carry out unsupervised classification of Swift-XRT time-series data using a variational autoencoder. We evaluate our model’s ability to identify different morphological classes by carrying out training on synthetic data. The generative aspect of the model can provide physical insight by highlighting the discriminative features in the light curves. We compare the classification results obtained with the traditional functional-form-based classification, and investigate the resulting level of segregation in the data set. We find that the data is segregated in the latent space according to traditional morphology. However, the observed gradual transition in between over-densities unifies the prevalent classification of GRBs based on their X-ray data into a single continuum, suggesting that light curves of different types should be unified under a single model.