Speaker
Description
The identification of gamma rays and suppression of cosmic-ray hadron background are crucial for very-high-energy gamma-ray observations and relevant scientific research of LHAASO-KM2A. Traditional machine learning methods, such as support vector machines, decision trees and deep neural networks have demonstrated promising performance in gamma-ray/hadron separation for ground-based experiments including H.E.S.S. and LHAASO KM2A. Recent advances in deep learning offer new opportunities for improved classification. In this work, we present a deep learning approach for gamma-ray/hadron separation, we model the simulated air showers as graph structures and analyze them through a dynamic graph convolutional neural network (DGCNN) framework. Our results demonstrate improvements in gamma-ray identification accuracy, particularly in the energy range above ~10 TeV, outperforming traditional methods employed by KM2A.
