Abstract
For a scintillating calorimeter, which is being designed to perform energy measurements in a next-generation high-energy collider experiment, a neural mapping is established to improve the overall detector performance. Training a neural network with energy vectors formed by the energy deposited on each cell of this granular detector, the original energy scale of the experimental particle beam is reconstructed and the linearity is significantly improved. In practice, the neural mapping corrects for nonlinearities that arise from the calorimeter design and it may replace classical methods that use energy dependent multiparameter functions.