Skip to Main content Skip to Navigation
Journal articles

The Tracking Machine Learning challenge : Throughput phase

Sabrina Amrouche 1 Laurent Basara 2 Paolo Calafiura 3 Dmitry Emeliyanov Victor Estrade 4 Steven Farrell 3 Cécile Germain 4 Vladimir Vava Gligorov 5 Tobias Golling 1 Sergey Gorbunov 6 Heather Gray 3 Isabelle Guyon 4 Mikhail Hushchyn 7 Vincenzo Innocente 1 Moritz Kiehn 1 Marcel Kunze 8 Edward Moyse 9 David Rousseau 10 Andreas Salzburger 1 Andrey Ustyuzhanin 7 Jean-Roch Vlimant 11
1 University of Geneva [Switzerland]
2 LAPP - Laboratoire d'Annecy de Physique des Particules
3 LBNL - Lawrence Berkeley National Laboratory [Berkeley]
4 LISN - Laboratoire Interdisciplinaire des Sciences du Numérique
5 LPNHE (UMR_7585) - Laboratoire de Physique Nucléaire et de Hautes Énergies
6 MIT - Massachusetts Institute of Technology
7 YSDA - Yandex School of Data Analysis
8 Heidelberg University
9 UMass Amherst - University of Massachusetts [Amherst]
10 IJCLab - Laboratoire de Physique des 2 Infinis Irène Joliot-Curie
11 CALTECH - California Institute of Technology
Abstract : This paper reports on the second "Throughput" phase of the Tracking Machine Learning (TrackML) challenge on the Codalab platform. As in the first "Accuracy" phase, the participants had to solve a difficult experimental problem linked to tracking accurately the trajectory of particles as e.g. created at the Large Hadron Collider (LHC): given (10^5) points, the participants had to connect them into (10^3) individual groups that represent the particle trajectories which are approximated helical. While in the first phase only the accuracy mattered, the goal of this second phase was a compromise between the accuracy and the speed of inference. Both were measured on the Codalab platform where the participants had to upload their software. The best three participants had solutions with good accuracy and speed an order of magnitude faster than the state of the art when the challenge was designed. Although the core algorithms were less diverse than in the first phase, a diversity of techniques have been used and are described in this paper. The performance of the algorithms are analysed in depth and lessons derived.
Keywords : Challenges
Document type :
Journal articles
Complete list of metadata
https://hal.inria.fr/hal-03159824
Contributor : Cecile Germain <>
Submitted on : Saturday, March 6, 2021 - 2:06:27 PM
Last modification on : Wednesday, April 14, 2021 - 3:40:56 AM

File

 Restricted access
To satisfy the distribution rights of the publisher, the document is embargoed until : 2021-09-06

Please log in to resquest access to the document

Identifiers

  • HAL Id : hal-03159824, version 2

Collections

IN2P3 | UNIV-SAVOIE | IJCLAB | CNRS | CENTRALESUPELEC | LPNHE | UP-SCIENCES | UNIV-PARIS | LAPP | LISN | SORBONNE-UNIVERSITE | SU-SCIENCES | SU-TI | LRI-AO | LISN-AO | GS-ENGINEERING

Citation

Sabrina Amrouche, Laurent Basara, Paolo Calafiura, Dmitry Emeliyanov, Victor Estrade, et al.. The Tracking Machine Learning challenge : Throughput phase. Computing and Software for Big Science, Springer, In press. ⟨hal-03159824v2⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

54

 

Contact                    Legal Notice                    Personal Data