Speaker
Description
An excess of gamma rays from the Galactic center is observed by the Fermi Space Telescope. The two leading hypotheses for the cause of this excess are millisecond pulsars or dark matter. Generically, we expect the statistics of these two sources to differ. We train a convolutional neural network (CNN) to accurately determine the relative flux contribution of point sources to the GCE, training the model on the energy dependent data for the first time. The CNN allows us to avoid biases that have been attributed to existing likelihood based techniques and we show training on energy dependent data produces results that are noticeably dimmer than those obtained by a CNN trained on energy independent data. This suggests there remains even further room for a dark matter contribution to the excess. We validate our results by testing on data simulated with known point source distributions and characterize the effects of mismodeling by testing the CNN on data generated with Galactic background models that differ from the model used to train the CNN.
