Open Collections

Abstract

The E36 Experiment at J-PARC was designed to obtain a precise measurement of the relative branching ratio Γ(K⁺ → e⁺v)/Γ(K⁺ → μ⁺v), using a stopped K⁺ beam. For this paper, a Convolutional Neural Network was used to analyze both the data obtained in the experimental and the simulated data. The purpose of this network is to identify instances where there is more than one particle track in the 256 multi-fibre kaon stopping target. These extra tracks may be produced when a positron collides with an electron (creating a delta ray). Identifying these extra tracks is crucial to calculating and adjusting the energy and momentum of the high energy lepton track leaving the target. Although the presented Neural Network provides very accurate results while classifying simulation data, its reliability for the experimental data remains unclear. Furthermore, its scope makes it too limited to replace the current methods used in this stage of the data analysis.