End-to-end multi-particle reconstruction in high occupancy imaging calorimeters with graph neural networks

Abstract

We present an end-to-end reconstruction algorithm to build particle candidates from detector hits in next-generation granular calorimeters similar to that foreseen for the high-luminosity upgrade of the CMS detector. The algorithm exploits a distance-weighted graph neural network, trained with object condensation, a graph segmentation technique. Through a single-shot approach, the reconstruction task is paired with energy regression. We describe the reconstruction performance in terms of efficiency as well as in terms of energy resolution. In addition, we show the jet reconstruction performance of our method and discuss its inference computational cost. To our knowledge, this work is the first-ever example of single-shot calorimetric reconstruction of $\mathcal{O}(1000)$ particles in high-luminosity conditions with 200 pileup.

References in corpus (20)

• The anti-k_t jet clustering algorithm
• An Introduction to PYTHIA 8.2
• Electron and photon reconstruction and identification with the CMS experiment at the CERN LHC
• A Roadmap for HEP Software and Computing R&D for the 2020s
• Learning representations of irregular particle-detector geometry with distance-weighted graph networks
• Calorimetry with Deep Learning: Particle Simulation and Reconstruction for Collider Physics
• Allen: A high level trigger on GPUs for LHCb
• MLPF: Efficient machine-learned particle-flow reconstruction using graph neural networks
• Charged particle tracking via edge-classifying interaction networks
• Distance-Weighted Graph Neural Networks on FPGAs for Real-Time Particle Reconstruction in High Energy Physics
• End-to-End Jet Classification of Quarks and Gluons with the CMS Open Data
• Object condensation: one-stage grid-free multi-object reconstruction in physics detectors, graph and image data
• Electromagnetic Showers Beyond Shower Shapes
• On the Use of Neural Networks for Energy Reconstruction in High-granularity Calorimeters
• Bringing heterogeneity to the CMS software framework
• A Comparison of CPU and GPU implementations for the LHCb Experiment Run 3 Trigger
• Usage of GPUs in ALICE Online and Offline processing during LHC Run 3
• Track Seeding and Labelling with Embedded-space Graph Neural Networks
• GPU-accelerated track reconstruction in the ALICE High Level Trigger
• Jet Single Shot Detection