Improving Energy Estimation at NOvA with Recurrent Neural Networks

Abstract

In this thesis I discuss the application of Recurrent Neural Networks to the problem of neutrino energy reconstruction at the NOvA experiment. NOvA is a long-baseline accelerator based neutrino oscillation experiment that holds one of the leading measurements of the $\Delta m_{32}^2$ oscillation parameter. In order to make precise measurements of the neutrino oscillation parameters, NOvA needs a good neutrino energy estimation algorithm. A new energy estimation algorithm that is based on a recurrent neural network architecture has been developed for NOvA. The new energy estimator has 15% better energy reconstruction than the previous energy estimation algorithm, and it is 5 times less sensitive to the major systematic uncertainty at NOvA. Using the new energy estimator has the potential to significantly improve the precision of measurements of the neutrino oscillation parameters at NOvA.