Cargando…

Graph Neural Network Track Reconstruction for the ATLAS ITk Detector

Graph Neural Networks (GNNs) have been shown to produce high accuracy performance on a variety of HEP tasks, including track reconstruction in the TrackML challenge, and tagging in jet physics. However, GNNs are less explored in applications with noisy, heterogeneous or ambiguous data. These element...

Descripción completa

Detalles Bibliográficos
Autores principales: Murnane, Daniel Thomas, Vallier, Alexis, Rougier, Charline, Calafiura, Paolo, Stark, Jan, Ju, Xiangyang, Farrell, Steven Andrew, Caillou, Sylvain, Neubauer, Mark, Atkinson, Markus Julian
Lenguaje:eng
Publicado: 2022
Materias:
Acceso en línea:http://cds.cern.ch/record/2809518
_version_ 1780973156678238208
author Murnane, Daniel Thomas
Vallier, Alexis
Rougier, Charline
Calafiura, Paolo
Stark, Jan
Ju, Xiangyang
Farrell, Steven Andrew
Caillou, Sylvain
Neubauer, Mark
Atkinson, Markus Julian
author_facet Murnane, Daniel Thomas
Vallier, Alexis
Rougier, Charline
Calafiura, Paolo
Stark, Jan
Ju, Xiangyang
Farrell, Steven Andrew
Caillou, Sylvain
Neubauer, Mark
Atkinson, Markus Julian
author_sort Murnane, Daniel Thomas
collection CERN
description Graph Neural Networks (GNNs) have been shown to produce high accuracy performance on a variety of HEP tasks, including track reconstruction in the TrackML challenge, and tagging in jet physics. However, GNNs are less explored in applications with noisy, heterogeneous or ambiguous data. These elements are expected from ATLAS Inner Tracker (ITk) detector data, when it is reformulated as a graph. We present the first comprehensive studies of a GNN-based track reconstruction pipeline on ATLAS-generated ITk data. Significant challenges exist in translating graph methods to this dataset. We analyze several approaches to low-latency and high-efficiency graph construction, including heuristics-based construction, discrete mappings of spacepoints to detector modules, and neural network learned mappings. We also extend these ideas to mappings of spacepoint doublets for more performant graph construction. Innovations in GNN training are required for ITk, and we discuss memory management for the very large ITk point clouds, and novel constructions of loss for noisy spacepoints and background tracks. Track candidates constructed from GNN link prediction may always suffer some inefficiency, particularly on noisy point clouds. We present several methods for post-processing GNN output for either very fast triplet seeding on GPU, or for recovering efficiency with learned embeddings of tracklets and with Kalman Filters. Finally, the performance of several configurations of GNN architecture based on the Interaction Network are considered, for various hardware and latency constraints.
id cern-2809518
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2022
record_format invenio
spelling cern-28095182022-05-18T06:19:47Zhttp://cds.cern.ch/record/2809518engMurnane, Daniel ThomasVallier, AlexisRougier, CharlineCalafiura, PaoloStark, JanJu, XiangyangFarrell, Steven AndrewCaillou, SylvainNeubauer, MarkAtkinson, Markus JulianGraph Neural Network Track Reconstruction for the ATLAS ITk DetectorParticle Physics - ExperimentGraph Neural Networks (GNNs) have been shown to produce high accuracy performance on a variety of HEP tasks, including track reconstruction in the TrackML challenge, and tagging in jet physics. However, GNNs are less explored in applications with noisy, heterogeneous or ambiguous data. These elements are expected from ATLAS Inner Tracker (ITk) detector data, when it is reformulated as a graph. We present the first comprehensive studies of a GNN-based track reconstruction pipeline on ATLAS-generated ITk data. Significant challenges exist in translating graph methods to this dataset. We analyze several approaches to low-latency and high-efficiency graph construction, including heuristics-based construction, discrete mappings of spacepoints to detector modules, and neural network learned mappings. We also extend these ideas to mappings of spacepoint doublets for more performant graph construction. Innovations in GNN training are required for ITk, and we discuss memory management for the very large ITk point clouds, and novel constructions of loss for noisy spacepoints and background tracks. Track candidates constructed from GNN link prediction may always suffer some inefficiency, particularly on noisy point clouds. We present several methods for post-processing GNN output for either very fast triplet seeding on GPU, or for recovering efficiency with learned embeddings of tracklets and with Kalman Filters. Finally, the performance of several configurations of GNN architecture based on the Interaction Network are considered, for various hardware and latency constraints.ATL-ITK-SLIDE-2022-119oai:cds.cern.ch:28095182022-05-16
spellingShingle Particle Physics - Experiment
Murnane, Daniel Thomas
Vallier, Alexis
Rougier, Charline
Calafiura, Paolo
Stark, Jan
Ju, Xiangyang
Farrell, Steven Andrew
Caillou, Sylvain
Neubauer, Mark
Atkinson, Markus Julian
Graph Neural Network Track Reconstruction for the ATLAS ITk Detector
title Graph Neural Network Track Reconstruction for the ATLAS ITk Detector
title_full Graph Neural Network Track Reconstruction for the ATLAS ITk Detector
title_fullStr Graph Neural Network Track Reconstruction for the ATLAS ITk Detector
title_full_unstemmed Graph Neural Network Track Reconstruction for the ATLAS ITk Detector
title_short Graph Neural Network Track Reconstruction for the ATLAS ITk Detector
title_sort graph neural network track reconstruction for the atlas itk detector
topic Particle Physics - Experiment
url http://cds.cern.ch/record/2809518
work_keys_str_mv AT murnanedanielthomas graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT vallieralexis graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT rougiercharline graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT calafiurapaolo graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT starkjan graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT juxiangyang graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT farrellstevenandrew graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT caillousylvain graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT neubauermark graphneuralnetworktrackreconstructionfortheatlasitkdetector
AT atkinsonmarkusjulian graphneuralnetworktrackreconstructionfortheatlasitkdetector