Track Finding in the LHCb Vertex Locator using Graph Neural Networks

Fuente: Zenodo
Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Correia, Anthony
Format: Recurso digital
Sprache:Englisch
Veröffentlicht: Zenodo 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866901434143866880
author Correia, Anthony
author_facet Correia, Anthony
contents <p>The start of Run 3 of the LHCb experiment in 2022 introduced Allen, a fully GPU-based first-level trigger system. This framework performs a near-complete reconstruction of proton-proton collisions at a rate of 30 MHz. Looking toward Run~5 and the High-Luminosity LHC (HL-LHC), increased luminosity and detector occupancy necessitate exploring alternative reconstruction methods, such as Artificial Intelligence, within this existing GPU infrastructure.</p> <p>This thesis presents the adaptation of the Exa.TrkX Graph Neural Network (GNN) pipeline, originally designed for central detectors like ATLAS and CMS, to the LHCb Vertex Locator (Velo). A key advantage of this study is the ability to benchmark the GNN pipeline against the standard "Search by triplet" algorithm on identical hardware.</p> <p>The adaptation and implementation rely on four main contributions: First, the preparation of large training and testing datasets by converting grid simulation files using the developed digout library. Second, an evaluation framework, MonteTracko, designed to compare the physics performance of the ETX4VELO pipeline against the baseline.<br>Third, architectural adaptations to address the specific challenges of a forward detector, such as well-defined sensor layers and crossing tracks. Fourth, the deployment of an inference pipeline in C++/CUDA within Allen. While training is performed in Python (PyTorch), production inference leverages TensorRT and custom CUDA kernels to handle combinatorial operations.</p> <p>The implemented pipeline achieves physics performance comparable to the Search by triplet algorithm. Although the current throughput is three orders of magnitude lower than the baseline, this work establishes a functional GNN tracking implementation on GPUs, identifying specific optimisations, such as quantisation, to address the speed gap.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_18421733
institution Zenodo
language eng
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle Track Finding in the LHCb Vertex Locator using Graph Neural Networks
Correia, Anthony
LHCb
Vertex Locator
Track Reconstruction
HL-LHC
Electron Reconstruction
Graph Neural Network
Exa.TrkX
GPU
Allen
CUDA/C++
TensorRT
ONNX
PyTorch
Real-Time Analysis
Quantisation
<p>The start of Run 3 of the LHCb experiment in 2022 introduced Allen, a fully GPU-based first-level trigger system. This framework performs a near-complete reconstruction of proton-proton collisions at a rate of 30 MHz. Looking toward Run~5 and the High-Luminosity LHC (HL-LHC), increased luminosity and detector occupancy necessitate exploring alternative reconstruction methods, such as Artificial Intelligence, within this existing GPU infrastructure.</p> <p>This thesis presents the adaptation of the Exa.TrkX Graph Neural Network (GNN) pipeline, originally designed for central detectors like ATLAS and CMS, to the LHCb Vertex Locator (Velo). A key advantage of this study is the ability to benchmark the GNN pipeline against the standard "Search by triplet" algorithm on identical hardware.</p> <p>The adaptation and implementation rely on four main contributions: First, the preparation of large training and testing datasets by converting grid simulation files using the developed digout library. Second, an evaluation framework, MonteTracko, designed to compare the physics performance of the ETX4VELO pipeline against the baseline.<br>Third, architectural adaptations to address the specific challenges of a forward detector, such as well-defined sensor layers and crossing tracks. Fourth, the deployment of an inference pipeline in C++/CUDA within Allen. While training is performed in Python (PyTorch), production inference leverages TensorRT and custom CUDA kernels to handle combinatorial operations.</p> <p>The implemented pipeline achieves physics performance comparable to the Search by triplet algorithm. Although the current throughput is three orders of magnitude lower than the baseline, this work establishes a functional GNN tracking implementation on GPUs, identifying specific optimisations, such as quantisation, to address the speed gap.</p>
title Track Finding in the LHCb Vertex Locator using Graph Neural Networks
topic LHCb
Vertex Locator
Track Reconstruction
HL-LHC
Electron Reconstruction
Graph Neural Network
Exa.TrkX
GPU
Allen
CUDA/C++
TensorRT
ONNX
PyTorch
Real-Time Analysis
Quantisation
url https://doi.org/10.5281/zenodo.18421733