Reconstruction of overlapping electromagnetic showers in calorimeters using Transformers

TL;DR

This paper introduces a novel deep learning approach using Transformers for reconstructing overlapping electromagnetic showers in calorimeters, significantly improving accuracy and robustness over traditional algorithms.

Contribution

The study presents ClusTEX, a single-stage graph transformer model with a new positional encoding scheme for efficient, geometry-aware clustering of calorimeter energy deposits.

Findings

Attention-based models outperform standard algorithms in overlapping shower reconstruction.

ClusTEX improves energy and position resolution in simulated calorimeter data.

Model maintains performance under detector failures and non-responsive channels.

Abstract

Accurate clustering of electromagnetic energy deposits is essential for reconstructing photons and electrons in modern hadron collider experiments, where boosted topologies and pileup cause overlapping showers and ambiguous energy assignment. We present deep learning-based clustering approaches that reconstruct particle energy and position directly from calorimeter readout. The study includes a two-step strategy in which candidate seed windows are identified and then jointly processed via distance-weighted message passing or attention mechanism and a single-step graph transformer, ClusTEX, which performs candidate selection and reconstruction in one inference stage. ClusTEX uses a novel positional encoding scheme that separates local coordinates within the graph from global detector coordinates, enabling efficient, geometry-aware inference. Models are trained on GEANT4 simulations of a simplified (toy) and an ECAL-inspired topology with an explicit $\eta-\phi$ dependence. Performance is evaluated using efficiency, energy and position resolutions and splitting rate - reconstruction of two objects for a single photon. In the toy calorimeter, attention-based interactions improve the reconstruction of overlapping showers relative to both the standard algorithm and distance-driven message passing, while maintaining performance on isolated photons and reducing splitting without multi-pass inference. For boosted $\pi^0\to\gamma\gamma$, the attention-based model retains di-photon mass reconstruction capability, where the standard algorithm becomes inefficient. In the ECAL-inspired topology, ClusTEX provides the best overall performance, yielding improved energy resolution and reduced splitting compared to two-step approaches and the standard algorithm. It also remains robust under localized detector failures, showing improved stability and partial recovery of energy in non-responsive channels.