Solving a differential equation for calibration of the ATLAS detector via a neural networks.
Calibration Example: using neural network solution for ODE-based calibration function
Supervisor : Pierre-Antoine Delsart, lecturer, ATLAS, LPSC, France.
The results of this internship have been presented in a workshop of the collaboration.
The particle physics studied at the LHC by the ATLAS detector requires high-precision measurements. precision. The energy quantities associated with hadronic jet objects cannot be used directly from experimental measurements, and must be corrected in order to be compared with theoretical predictions. from experimental measurements, and must be corrected before they can be compared with theoretical predictions. This addresses this problem in the form of a nonlinear second-order differential equation on the calibration function function, which has no analytical solution. A neural network is used to solve it. The aim is to is to obtain an accurate calibration function by exploiting the non-linear capabilities of neural networks.
