Speaker
Description
The Southern Wide-field Gamma-ray Observatory (SWGO) is a next-generation ground-based gamma-ray observatory. Currently in the R&D phase, the experiment is expected to have a large array of water Cherenkov detectors (WCD) placed at high elevations in South America. This will enable precise observations of the gamma-ray sky, mainly in the regime of ~100GeV up to the PeV region. The primary background of gamma-ray observations are hadronic showers, that need to be rejected to guarantee a high signal-to-noise ratio.
Currently approaches heavily rely on hand-crafted features or employ a large number of variables which are exploited using machine learning techniques. In this talk I will discuss a novel approach for gamma/hadron separation for SWGO. Based on the great success of deep learning in engineering and science, we present an innovative classification algorithm that processes lowlevel information at station level using graph neural networks. We will examine the performance of the novel technique and compare it to a machine learning algorithm that is currently in use.
