-
TrackFormers: In Search of Transformer-Based Particle Tracking for the High-Luminosity LHC Era
Authors:
Sascha Caron,
Nadezhda Dobreva,
Antonio Ferrer Sánchez,
José D. Martín-Guerrero,
Uraz Odyurt,
Roberto Ruiz de Austri Bazan,
Zef Wolffs,
Yue Zhao
Abstract:
High-Energy Physics experiments are facing a multi-fold data increase with every new iteration. This is certainly the case for the upcoming High-Luminosity LHC upgrade. Such increased data processing requirements forces revisions to almost every step of the data processing pipeline. One such step in need of an overhaul is the task of particle track reconstruction, a.k.a., tracking. A Machine Learn…
▽ More
High-Energy Physics experiments are facing a multi-fold data increase with every new iteration. This is certainly the case for the upcoming High-Luminosity LHC upgrade. Such increased data processing requirements forces revisions to almost every step of the data processing pipeline. One such step in need of an overhaul is the task of particle track reconstruction, a.k.a., tracking. A Machine Learning-assisted solution is expected to provide significant improvements, since the most time-consuming step in tracking is the assignment of hits to particles or track candidates. This is the topic of this paper.
We take inspiration from large language models. As such, we consider two approaches: the prediction of the next word in a sentence (next hit point in a track), as well as the one-shot prediction of all hits within an event. In an extensive design effort, we have experimented with three models based on the Transformer architecture and one model based on the U-Net architecture, performing track association predictions for collision event hit points. In our evaluation, we consider a spectrum of simple to complex representations of the problem, eliminating designs with lower metrics early on. We report extensive results, covering both prediction accuracy (score) and computational performance. We have made use of the REDVID simulation framework, as well as reductions applied to the TrackML data set, to compose five data sets from simple to complex, for our experiments. The results highlight distinct advantages among different designs in terms of prediction accuracy and computational performance, demonstrating the efficiency of our methodology. Most importantly, the results show the viability of a one-shot encoder-classifier based Transformer solution as a practical approach for the task of tracking.
△ Less
Submitted 9 July, 2024;
originally announced July 2024.
-
Novel Approaches for ML-Assisted Particle Track Reconstruction and Hit Clustering
Authors:
Uraz Odyurt,
Nadezhda Dobreva,
Zef Wolffs,
Yue Zhao,
Antonio Ferrer Sánchez,
Roberto Ruiz de Austri Bazan,
José D. Martín-Guerrero,
Ana-Lucia Varbanescu,
Sascha Caron
Abstract:
Track reconstruction is a vital aspect of High-Energy Physics (HEP) and plays a critical role in major experiments. In this study, we delve into unexplored avenues for particle track reconstruction and hit clustering. Firstly, we enhance the algorithmic design effort by utilising a simplified simulator (REDVID) to generate training data that is specifically composed for simplicity. We demonstrate…
▽ More
Track reconstruction is a vital aspect of High-Energy Physics (HEP) and plays a critical role in major experiments. In this study, we delve into unexplored avenues for particle track reconstruction and hit clustering. Firstly, we enhance the algorithmic design effort by utilising a simplified simulator (REDVID) to generate training data that is specifically composed for simplicity. We demonstrate the effectiveness of this data in guiding the development of optimal network architectures. Additionally, we investigate the application of image segmentation networks for this task, exploring their potential for accurate track reconstruction. Moreover, we approach the task from a different perspective by treating it as a hit sequence to track sequence translation problem. Specifically, we explore the utilisation of Transformer architectures for tracking purposes. Our preliminary findings are covered in detail. By considering this novel approach, we aim to uncover new insights and potential advancements in track reconstruction. This research sheds light on previously unexplored methods and provides valuable insights for the field of particle track reconstruction and hit clustering in HEP.
△ Less
Submitted 27 May, 2024;
originally announced May 2024.
-
LHC bounds on MUED after Run 2
Authors:
Marvin M. Flores,
Jong Soo Kim,
Krzysztof Rolbiecki,
Roberto Ruiz de Austri Bazan
Abstract:
We present updated LHC limits on the minimal universal extra dimensions (MUED) model from the Run 2 searches. We scan the parameter space against a number of searches implemented in the public code \textsc{CheckMATE} and derive up-to-date limits on the MUED parameter space from 13 TeV searches. The strongest constraints come from a search dedicated to squarks and gluinos with one isolated lepton,…
▽ More
We present updated LHC limits on the minimal universal extra dimensions (MUED) model from the Run 2 searches. We scan the parameter space against a number of searches implemented in the public code \textsc{CheckMATE} and derive up-to-date limits on the MUED parameter space from 13 TeV searches. The strongest constraints come from a search dedicated to squarks and gluinos with one isolated lepton, jets and missing transverse energy. In the procedure we take into account initial state radiation and stress its importance in the MUED searches, which is not always appreciated.
△ Less
Submitted 14 August, 2022; v1 submitted 1 October, 2021;
originally announced October 2021.
-
Constraining electroweak and strongly charged long-lived particles with CheckMATE
Authors:
Nishita Desai,
Florian Domingo,
Jong Soo Kim,
Roberto Ruiz de Austri Bazan,
Krzysztof Rolbiecki,
Mangesh Sonawane,
Zeren Simon Wang
Abstract:
Long-lived particles have become a new frontier in the exploration of physics beyond the Standard Model. In this paper, we present the implementation of four types of long-lived particle searches, viz. displaced leptons, disappearing track, displaced vertex (together with muons or with missing energy), and heavy charged tracks. These four categories cover the signatures of a large range of physics…
▽ More
Long-lived particles have become a new frontier in the exploration of physics beyond the Standard Model. In this paper, we present the implementation of four types of long-lived particle searches, viz. displaced leptons, disappearing track, displaced vertex (together with muons or with missing energy), and heavy charged tracks. These four categories cover the signatures of a large range of physics models. We illustrate their potential for exclusion and discuss their mutual overlaps in mass-lifetime space for two simple phenomenological models involving either a U(1)-charged or a coloured scalar.
△ Less
Submitted 5 November, 2021; v1 submitted 9 April, 2021;
originally announced April 2021.
-
Multi-Time Systems of Conservation Laws
Authors:
Aldo Bazan,
Paola Loreti,
Wladimir Neves
Abstract:
Motivated by the work of P.L. Lions and J-C. Rochet [12], concerning multi-time Hamilton-Jacobi equations, we introduce the theory of multi-time systems of conservation laws. We show the existence and uniqueness of solution to the Cauchy problem for a system of multi-time conservation laws with two independent time variables in one space dimension. Our proof relies on a suitable generalization of…
▽ More
Motivated by the work of P.L. Lions and J-C. Rochet [12], concerning multi-time Hamilton-Jacobi equations, we introduce the theory of multi-time systems of conservation laws. We show the existence and uniqueness of solution to the Cauchy problem for a system of multi-time conservation laws with two independent time variables in one space dimension. Our proof relies on a suitable generalization of the Lax-Oleinik formula.
△ Less
Submitted 27 May, 2013;
originally announced May 2013.
-
A scaling approach to Caffarelli-Kohn-Nirenberg inequality
Authors:
Aldo Bazan,
Wladimir Neves
Abstract:
We mainly consider the general Caffarelli-Kohn-Nirenberg inequality in the Euclidean and Riemannian setting. In both cases, our proof relies mostly on a new parameter s conveniently introduced, see (2.7).
We mainly consider the general Caffarelli-Kohn-Nirenberg inequality in the Euclidean and Riemannian setting. In both cases, our proof relies mostly on a new parameter s conveniently introduced, see (2.7).
△ Less
Submitted 3 April, 2014; v1 submitted 5 April, 2013;
originally announced April 2013.
-
The Hardy and Caffarelli-Kohn-Nirenberg Inequalities Revisited
Authors:
Aldo Bazan,
Wladimir Neves
Abstract:
In this paper some important inequalities are revisited. First, as motivation, we give another proof of the Hardy's inequality applying convenient vector fields as introduced by Mitidieri, see [6]. Then, we investigate a particular case of the Caffarelli-Kohn-Nirenberg's inequality. Finally, we study the Rellic's inequality.
In this paper some important inequalities are revisited. First, as motivation, we give another proof of the Hardy's inequality applying convenient vector fields as introduced by Mitidieri, see [6]. Then, we investigate a particular case of the Caffarelli-Kohn-Nirenberg's inequality. Finally, we study the Rellic's inequality.
△ Less
Submitted 12 July, 2010;
originally announced July 2010.
-
Conformal map** of ultrasonic crystals: confining ultrasound and cochlear-like wave guiding
Authors:
A. Bazan,
M. Torres,
F. R. Montero de Espinosa,
R. Quintero-Torres,
J. L. Aragon
Abstract:
Conformal map** of a slab of a two-dimensional ultrasonic crystal generate a closed geometrical arrangement of ultrasonic scatterers with appealing acoustic properties. This acoustic shell is able to confine ultrasonic modes. Some of these internal resonances can be induced from an external wave source. The map** of a linear defect produces a wave-guide that exhibits a spatial-frequency sele…
▽ More
Conformal map** of a slab of a two-dimensional ultrasonic crystal generate a closed geometrical arrangement of ultrasonic scatterers with appealing acoustic properties. This acoustic shell is able to confine ultrasonic modes. Some of these internal resonances can be induced from an external wave source. The map** of a linear defect produces a wave-guide that exhibits a spatial-frequency selection analogous to that characteristic of a synthetic "cochlea". Both, experimental and theoretical results are reported here.
△ Less
Submitted 25 January, 2007;
originally announced January 2007.
-
Quasicrystalline and rational approximant wave patterns in hydrodynamic and quantum nested wells
Authors:
A. Bazán,
M. Torres,
G. Chiappe,
E. Louis,
J. A. Miralles,
J. A. Vergés,
Gerardo G. Naumis,
J. L. Aragón
Abstract:
The eigenfunctions of nested wells with incommensurate boundary geometry, in both hydrodynamic shallow water regime and quantum cases, are systematically and exhaustively studied in this letter. The boundary arrangement of the nested wells consist of polygonal ones, square or hexagonal, with a concentric immersed similar but rotated well or plateau. A rich taxonomy of wave patterns, such as quas…
▽ More
The eigenfunctions of nested wells with incommensurate boundary geometry, in both hydrodynamic shallow water regime and quantum cases, are systematically and exhaustively studied in this letter. The boundary arrangement of the nested wells consist of polygonal ones, square or hexagonal, with a concentric immersed similar but rotated well or plateau. A rich taxonomy of wave patterns, such as quasicrystalline states, their crystalline rational approximants and some other exotic but well known tilings, is found in these mimicked experiments. To our best knowledge, these hydrodynamic rational approximants are presented here for the first time in a hydrodynamic-quantum framework. The corresponding statistical nature of the energy level spacing distribution reflects this taxonomy by changing the spectral types.
△ Less
Submitted 2 July, 2006;
originally announced July 2006.
-
The Athena Data Dictionary and Description Language
Authors:
Alain Bazan,
Thierry Bouedo,
Philippe Ghez,
Massimo Marino,
Craig Tull
Abstract:
Athena is the ATLAS off-line software framework, based upon the GAUDI architecture from LHCb. As part of ATLAS' continuing efforts to enhance and customise the architecture to meet our needs, we have developed a data object description tool suite and service for Athena. The aim is to provide a set of tools to describe, manage, integrate and use the Event Data Model at a design level according to…
▽ More
Athena is the ATLAS off-line software framework, based upon the GAUDI architecture from LHCb. As part of ATLAS' continuing efforts to enhance and customise the architecture to meet our needs, we have developed a data object description tool suite and service for Athena. The aim is to provide a set of tools to describe, manage, integrate and use the Event Data Model at a design level according to the concepts of the Athena framework (use of patterns, relationships, ...). Moreover, to ensure stability and reusability this must be fully independent from the implementation details. After an extensive investigation into the many options, we have developed a language grammar based upon a description language (IDL, ODL) to provide support for object integration in Athena. We have then developed a compiler front end based upon this language grammar, JavaCC, and a Java Reflection API-like interface. We have then used these tools to develop several compiler back ends which meet specific needs in ATLAS such as automatic generation of object converters, and data object scripting interfaces. We present here details of our work and experience to date on the Athena Definition Language and Athena Data Dictionary.
△ Less
Submitted 28 May, 2003;
originally announced May 2003.
-
A data capture and production management tool for the assembly and construction of the CMS detectors
Authors:
J-M Le Goff,
J-P Vialle,
A Bazan,
T Le Flour,
S Lieunard,
D Rousset,
R McClatchey,
N Baker,
H Heath,
Z Kovacs,
E Leonardi,
G Barone,
G Organtini
Abstract:
The CMS experiment will comprise several very large high resolution detectors for physics. Each detector may be constructed of well over a million parts and will be produced and assembled during the next decade by specialised centres distributed world-wide. Each constituent part of each detector must be accurately measured and tested locally prior to its ultimate assembly and integration in the…
▽ More
The CMS experiment will comprise several very large high resolution detectors for physics. Each detector may be constructed of well over a million parts and will be produced and assembled during the next decade by specialised centres distributed world-wide. Each constituent part of each detector must be accurately measured and tested locally prior to its ultimate assembly and integration in the experimental area at CERN. The CRISTAL project (Concurrent Repository and Information System for Tracking Assembly and production Lifecycles) [1] aims to monitor and control the quality of the production and assembly process to aid in optimising the performance of the physics detectors and to reject unacceptable constituent parts as early as possible in the construction lifecycle. During assembly CRISTAL will capture all the information required for subsequent detector calibration. Distributed instances of Object databases linked via CORBA [2] and with WWW/Java-based query processing are the main technology aspects of CRISTAL.
△ Less
Submitted 18 May, 2001;
originally announced May 2001.
-
Detector Construction Management and Quality Control: Establishing and Using a CRISTAL System
Authors:
J-M. Le Goff,
G. Chevenier,
A. Bazan,
T. Le Flour,
S. Lieunard,
S. Murray,
J-P. Vialle,
N. Baker,
F. Estrella,
Z. Kovacs,
R. McClatchey,
G. Organtini,
S. Bityukov
Abstract:
The CRISTAL (Cooperating Repositories and an Information System for Tracking Assembly Lifecycles) project is delivering a software system to facilitate the management of the engineering data collected at each stage of production of CMS. CRISTAL captures all the physical characteristics of CMS components as each sub-detector is tested and assembled. These data are retained for later use in areas…
▽ More
The CRISTAL (Cooperating Repositories and an Information System for Tracking Assembly Lifecycles) project is delivering a software system to facilitate the management of the engineering data collected at each stage of production of CMS. CRISTAL captures all the physical characteristics of CMS components as each sub-detector is tested and assembled. These data are retained for later use in areas such as detector slow control, calibration and maintenance. CRISTAL must, therefore, support different views onto its data dependent on the role of the user. These data viewpoints are investigated in this paper. In the recent past two CMS Notes have been written about CRISTAL. The first note, CMS 1996/003, detailed the requirements for CRISTAL, its relationship to other CMS software, its objectives and reviewed the technology on which it would be based. CMS 1997/104 explained some important design concepts on which CRISTAL is and showed how CRISTAL integrated the domains of product data man- agement and workflow management. This note explains, through the use of diagrams, how CRISTAL can be established for detector production and used as the information source for analyses, such as calibration and slow controls, carried out by physicists. The reader should consult the earlier CMS Notes and conference papers for technical detail on CRISTAL - this note concentrates on issues surrounding the practical use of the CRISTAL software.
△ Less
Submitted 8 June, 1998;
originally announced June 1998.
-
From Design to Production Control Through the Integration of Engineering Data Management and Workflow Management Systems
Authors:
J-M. Le Goff,
G. Chevenier,
A. Bazan,
T. Le Flour,
S. Lieunard,
S. Murray,
J-P. Vialle,
N. Baker,
F. Estrella,
Z. Kovacs,
R. McClatchey,
G. Organtini,
S. Bityukov
Abstract:
At a time when many companies are under pressure to reduce "times-to-market" the management of product information from the early stages of design through assembly to manufacture and production has become increasingly important. Similarly in the construction of high energy physics devices the collection of (often evolving) engineering data is central to the subsequent physics analysis. Tradition…
▽ More
At a time when many companies are under pressure to reduce "times-to-market" the management of product information from the early stages of design through assembly to manufacture and production has become increasingly important. Similarly in the construction of high energy physics devices the collection of (often evolving) engineering data is central to the subsequent physics analysis. Traditionally in industry design engineers have employed Engineering Data Management Systems (also called Product Data Management Systems) to coordinate and control access to documented versions of product designs. However, these systems provide control only at the collaborative design level and are seldom used beyond design. Workflow management systems, on the other hand, are employed in industry to coordinate and support the more complex and repeatable work processes of the production environment. Commercial workflow products cannot support the highly dynamic activities found both in the design stages of product development and in rapidly evolving workflow definitions. The integration of Product Data Management with Workflow Management can provide support for product development from initial CAD/CAM collaborative design through to the support and optimisation of production workflow activities. This paper investigates this integration and proposes a philosophy for the support of product data throughout the full development and production lifecycle and demonstrates its usefulness in the construction of CMS detectors.
△ Less
Submitted 27 February, 1998; v1 submitted 6 February, 1998;
originally announced February 1998.
