4 sujets /DPhN/LSN

Dernière mise à jour :


 

MACHINE LEARNING FOR INVERSE PROBLEMS IN HADRON STRUCTURE

SL-DRF-24-0306

Research field : Nuclear Physics
Location :

Service de Physique Nucléaire (DPhN)

Laboratoire structure du nucléon (LSN) (LSN)

Saclay

Contact :

Valerio Bertone

Hervé Moutarde

Starting date : 01-10-2024

Contact :

Valerio Bertone
CEA - DRF/IRFU/DPhN/LSN


Thesis supervisor :

Hervé Moutarde
CEA - DRF/IRFU/DPhN

33 1 69 08 32 06

Laboratory link : https://irfu.cea.fr/Phocea/Vie_des_labos/Ast/ast_groupe.php?id_groupe=4189

Characterizing the multidimensional structure of hadrons in terms of quarks and gluons is one of the major objectives of hadronic physics today. This is not only the central theme of many experimental facilities worldwide, but also one of the main reasons for the construction of future colliders in the USA and China. It is also one of the key areas of research for intensive numerical simulations of the strong interaction. However, in both cases, the connection between measured and simulated data on the one hand, and the multidimensional structure of hadrons on the other, is not direct. The data are linked to the hadron structure via mathematically ill-posed multidimensional inverse problems. It has been shown that these inverse problems lead to a significant increase in uncertainties, to the point of becoming the dominant source of uncertainty in some cases. The aim of this thesis is to use machine learning tools to assess, reduce and correctly propagate uncertainties from experimental or simulation data to the multidimensional structure of hadrons. The strategy for achieving this is to develop an original neural network architecture capable of taking into account the full range of theoretical properties arising from quantum chromodynamics, and then to adapt it to inverse problems linking experimental and simulation data to the 3D structure of hadrons.
Study of the first Xenon-136 double-beta decay events of the PandaX-III experiment with neural network techniques

SL-DRF-24-0392

Research field : Particle physics
Location :

Service de Physique Nucléaire (DPhN)

Laboratoire structure du nucléon (LSN) (LSN)

Saclay

Contact :

Damien NEYRET

Starting date : 01-10-2024

Contact :

Damien NEYRET
CEA - DRF/IRFU/DPhN/LSN

01 69 08 75 52

Thesis supervisor :

Damien NEYRET
CEA - DRF/IRFU/DPhN/LSN

01 69 08 75 52

More : https://pandax.sjtu.edu.cn/

The PandaX-III collaboration proposes to determine whether the neutrino is a Majorana particle, i.e. its own antiparticle. In this purpose this international collaboration, in which the Research Institute on Fundamental laws of the Universe (IRFU) of CEA Saclay participates, aims to observe neutrinoless double-beta decays of the Xenon-136, where the emission of the two electrons is not compensated by the simultaneous emission of two anti-neutrinos. Such an observation would violate the principle of conservation of leptonic number, in opposition with the predictions of the standard model of particle physics. The search of such rare events requires an enormous quantity of Xenon-136 atoms, a deep underground laboratory protected from cosmic rays and with low radioactive levels, like the Jinping underground laboratory (CJPL, Sichuan province, China), and a very effective particle detector.

The first phase of the experiment aims to construct a first TPC module (Time Projection Chamber) of 145kg of Xenon-136, which will be followed in a second stage by four other 200kg modules. The TPC will be equipped with detectors able to measure the energy of the two beta electrons with an excellent accuracy. The first TPC module will be commissioned end of 2024. The trajectory of the two electrons emitted by the double-beta decay will be reconstructed to measure the initial energy of those electrons, and to recognize the topology of their trajectories to differentiate them from gamma backgrounds which emit only one electron. That module will be equipped with gaseous Micromegas detectors which have a good energy resolution and a very good radio-purity which limits the amount of gamma backgrounds coming from radioactive contamination.

The PandaX-III collaboration is working on the construction of the first TPC module. It will be installed at CJPL during the year 2024. Reconstruction algorithms of detector data using neural networks are being developed, in order to complete the analytical methods already implemented in the REST environment of data reconstruction and analysis, to optimize double-beta events versus gamma backgrounds discrimination, and to improve the quality of the electron energy reconstruction. These algorithms are trained and evaluated on simulated Monte-Carlo events. Data from reduced-size TPC prototype will be also used to test these algorithms in real conditions. As soon as the first module will be installed end of 2024 these algorithms will be used for detector calibrations and for being implemented in real data analysis. They will be then used to extract the first physics results on double-beta events.

The main task of the PhD student will be to contribute on the development of data reconstruction algorithms based on neural networks, in particular by taking into account the defects of the detectors (dead channels, performance inhomogeneity, gas impurities, etc...) and by implementing in REST the data correction methods needed to compensate these defects. That work will include studies of data from prototype TPC chambers, as well as Monte-Carlo simulations. Moreover, as soon as the data from the first TPC module will be available the student will participate to the data analysis and the extraction of the physics results. These studies will be presented in conferences and published in scientific journals. The student will also participate to an R&D to optimize Micromegas detectors in order to improve their energy resolution as well as their general performance in high pressure gaseous Xenon.

A Master internship of 4 to 6 months would be also possible in the IRFU/DPhN PandaX-III group before the start of the PhD thesis.
Lambda hyperon polarization measurement in exclusive deeply virtual meson production processes

SL-DRF-24-0386

Research field : Particle physics
Location :

Service de Physique Nucléaire (DPhN)

Laboratoire structure du nucléon (LSN) (LSN)

Saclay

Contact :

Francesco BOSSU

Starting date : 01-10-2024

Contact :

Francesco BOSSU
CEA - DRF/IRFU/SPhN


Thesis supervisor :

Francesco BOSSU
CEA - DRF/IRFU/SPhN


This thesis focuses on measuring the polarization of Lambda hyperons in exclusive deeply virtual meson production processes. The study is rooted in a surprising discovery from the 1970s: in proton-Beryllium collisions, ? hyperons exhibited transverse polarization, challenging the predictions of perturbative Quantum Chromodynamics. Similar polarizations have since been observed in various collision systems.
The proposed research topic leverages deeply virtual exclusive reactions in electron-proton scattering, providing precise control over final states and initial particle polarizations. Specifically, the reaction e+p->e+Lambda+K+ is explored to shed light on the Lambda hyperon's polarization. This process is also sensitive to the poorly known transversity Generalized Parton Distributions (GPDs) of the nucleon, offering valuable insights into nucleon properties.
The thesis aims to analyze data collected with the CLAS12 experiment at the Jefferson Laboratory (JLab) in US, with a focus on e-p collisions with a longitudinally polarized NH3 target. Machine learning algorithms and simulations will be employed to enhance data reconstruction and event candidate selection. The candidate will also contribute to simulation studies for future detectors and their reconstruction algorithms for the EIC.
The research will be conducted within the Laboratory of Nucleon Structure at CEA/Irfu. A background in particle physics, computer science (C++, Python), and knowledge of particle detectors is beneficial for active participation in data analysis.
The student will have the opportunity to collaborate with local and international researchers, to participate in the CLAS collaboration, to join the EIC user group with frequent trips to the USA for data collection and workshops, and present research findings at international conferences.
Accessing the 3D structure of pions with CLAS12

SL-DRF-24-0328

Research field : Particle physics
Location :

Service de Physique Nucléaire (DPhN)

Laboratoire structure du nucléon (LSN) (LSN)

Saclay

Contact :

Maxime DEFURNE

Damien NEYRET

Starting date : 01-10-2024

Contact :

Maxime DEFURNE
CEA - DRF/IRFU/DPhN/LSN

01 69 08 32 37

Thesis supervisor :

Damien NEYRET
CEA - DRF/IRFU/DPhN/LSN

01 69 08 75 52

Laboratory link : https://irfu.cea.fr/dphn/

In collaboration with the Thomas Jefferson Laboratory (JLab) in the USA, the researchers in the laboratory of nucleon structure at Irfu want to understand how quarks and gluons interact to form hadrons such protons, neutrons and pions. At JLab, a 11-GeV electron beam is impinged on a proton target. The protos are constituted of three quarks surrounded by a cloud of quark/antiquark pairs whose quantum numbers are similar to pions. The electrons of the beam will interact with these pairs with a structure analogous to a pion. More specifically, we are interested in the deeply virtual Compton scattering (DVCS) giving correlations between longitudinal momentum and transverse position of quarks in a pion. In other words, we are going to perform the very first 3D study of the pion structure. The PhD student will analyze data already available to isolate the DVCS events. A digital twins of the Monte-Carlo simulation/reconstruction chain will be produced with a conditional Generative Adversarial Network in order to caracterize faster and more accurately the background and, in the end, subtract it. The PhD student will travel two to three times a year to JLab, participating to the data taking as well as attending the collaboration meeting. The results will be presented in international conferences and published in peer-reviewed journals.

• Nuclear Physics

• Particle physics

 

Retour en haut