High Granularity Calorimeter Reconstruction Results using a Graph Neural Network

We present the current stage of research progress towards a one-pass, completely Machine Learning (ML) based reconstruction. The model used is based on graph Neural Networks (GNNs) and analyzes the hits in each HGCAL endcap. The ML algorithm is trained to predict clusters of hits originating from the same incident particle by labelling the hits with the same cluster index. We impose simple criteria to assess whether the hits associated as a cluster by the prediction are matched to any of clusters of hits resulting from individual incident particles. While this algorithm is not yet ready for the 200 PU target of the HL-LHC, it is nevertheless useful to show its performance on a physically interesting dataset beyond single particle reconstruction. For this purpose the algorithm is studied by pointing two tau leptons each at both HGCAL endcaps, where each tau may decay according to its measured standard model branching probabilities. This includes the material interaction of the tau decay products which may create additional particles incident upon the calorimeter. Using this varied multiparticle environment we can investigate the application of this reconstruction technique and begin to characterize energy containment and performance of this reconstruction in these more complex environments.