Data Representation Optimization for ML-based calorimeter simulation

This project aims to develop a pipeline for producing calorimeter shower datasets in point-cloud form for a given detector and various types of particles. Its central goal is not only to compress detailed Geant4 shower information into a compact point-cloud representation, but to do so while preserving the physics content needed for the downstream. The work will proceed in two major stages. First, electromagnetic (EM) showers will be used to establish the pipeline, optimize point-cloud construction, and define a robust validation protocol. Second, the same framework will be extended to hadronic showers, which are more challenging due to larger fluctuations, invisible energy, delayed components, and a stronger dependence on material and interaction modeling.