Automatic Design of Artificial Neural Networks for Gamma-Ray Detection

The goal of this work is to investigate the possibility of improving current gamma/hadron discrimination based on the shower patterns recorded on the ground. To this end, we propose the use of Convolutional Neural Networks (CNNs) for their ability to distinguish patterns based on automatically designed features. In order to promote the creation of CNNs that accurately uncover the hidden patterns in the data, and at the same time avoid the burden of hand-crafting the topology and learning hyper-parameters we resort to NeuroEvolution; in particular, we use Fast-DENSER++, a variant of Deep Evolutionary Network Structured Representation. The results show that the best CNN generated by Fast-DENSER++ improves by a factor of 2.37 when compared to the results reported by classical statistical approaches. Additionally, we experiment ensembling the best-generated CNNs; the ensemble leads to an improvement by a factor of 2.48. These results establish a new state-of-the-art in the gamma/hadron discrimination problem, based on the ground impact patterns, and thus prove that CNNs automatically discovered by Fast-DENSER++ can be used to enable investment savings due to the need for smaller grids of sensors.