HAhRD: DeepReconstruction

One of the challenges faced in Particle Physics Experiment after the collision of particles in LHC is the reconstruction of the events.This includes finding the type of daughter particles created and other important characteristics associated with particles like energy, from the data recorded by Detectors like CMS or ATLAS.

This project is targeted on event reconstruction of particles produced after the proton-proton collision, from data recorded in one of future sub-detector of CMS named as HGCAL(High Granularity Calorimeter). We will be using CNN (Convolutional Neural Network) for reconstructing the rare processes by classifying and learning other characteristics of the particles from the hits (energy deposits) recorded in the detector which are generated after the collision.

The main goal of this project is to develop a software pipeline, compatible with HGCAL sub-detector, which can be used by Physicist or other developers to create and train a CNN architecture on GPU to get additional insights in event reconstruction.