PhD subjects

16 sujets IRFU/DAP

Dernière mise à jour : 14-04-2021


• Artificial intelligence & Data intelligence

• Astrophysics

• Instrumentation

 

"3x2pt" analysis: Cross-correlations of cosmological probes, and application to state-of-the-art weak-lensing and galaxy clustering surveys

SL-DRF-21-0278

Location :

Direction d’Astrophysique (DAP)

Laboratoire CosmoStat (LCS)

Saclay

Contact :

Martin Kilbinger

Valeria Pettorino

Starting date : 01-09-2021

Contact :

Martin Kilbinger
CEA - DRF/IRFU/DAP/LCS

21753

Thesis supervisor :

Valeria Pettorino
CEA - DRF/IRFU/DAP/LCS


Personal web page : http://www.cosmostat.org/jobs/mk_3x2pt_2020

Laboratory link : http://www.cosmostat.org/

The large upcoming cosmological experiments such as the space telescopes Euclid and Nancy Grace Roman, or the ground-based survey LSST, will use two main probes with the goal to measure the properties of dark matter and dark energy: weak gravitational lensing, which is the deformation of the images of distant galaxies by tidal fields on very large scales, and galaxy clustering, denoting the distribution of galaxies in the cosmic web.



This PhD thesis will explore the cross-correlation between those two probes, which is of great importance since these observables are sensitive to the same structures, which interrelates them. This work will be applied to data from the surveys UNIONS and DESI. The student will be trained at the interface between theory and observations, and provide key tools for upcoming large surveys.
Weak-gravitational lensing mass maps for cosmology and gravitational wave astronomy

SL-DRF-21-0350

Location :

Direction d’Astrophysique (DAP)

Laboratoire CosmoStat (LCS)

Saclay

Contact :

Martin Kilbinger

Starting date : 01-09-2021

Contact :

Martin Kilbinger
CEA - DRF/IRFU/DAP/LCS

21753

Thesis supervisor :

Martin Kilbinger
CEA - DRF/IRFU/DAP/LCS

21753

Personal web page : http://www.cosmostat.org/jobs/mk_wlcosmogw_2020

Laboratory link : http://www.cosmostat.org

Weak lensing denotes galaxy image distortions induced by structures on large scales. Dark-matter maps obtained from weak lensing help us to shed light on the mystery of the recent acceleration of the Universe. In addition, they are important for gravitational waves (GW), which can be magnified by dark matter along the line of sight.



This PhD thesis will develop methods to analyse weak-lensing data. Machine-learning techniques will be used to calibrate the measurements. These methods will be applied to survey UNIONS (Ultraviolet Near-Infrared Optical Northen Sky Survey), and WFST (Wide-Field Survey Telescope), which will be built in China. The goal is to measure the properties of dark energy, and to correct magnification of GW signals.
Uncertainties for large scale deep learning-based image reconstruction

SL-DRF-21-0336

Research field : Artificial intelligence & Data intelligence
Location :

Direction d’Astrophysique (DAP)

Laboratoire CosmoStat (LCS)

Saclay

Contact :

Jean-Luc STARCK

Starting date : 01-10-2021

Contact :

Jean-Luc STARCK
CEA - DSM/IRFU/SAp/LCS

01 69 08 57 64

Thesis supervisor :

Jean-Luc STARCK
CEA - DSM/IRFU/SAp/LCS

01 69 08 57 64

Personal web page : http://jstarck.cosmostat.org

Laboratory link : http://www.cosmostat.org

Deep learning (DL) has changed the way of solving inverse problems. Many scientific challenges remain that must be met for its deployment in astronomical imagery: i) taking into account the physical training model, ii) estimating the uncertainties on reconstructed images, iii) generalization, and iv ) the volume of data for scaling up. To quantify the uncertainties, we have introduced a probabilistic DL approach (Remy et al., 2020), which makes it possible to derive the a posteriori distribution of the solution. This requires however to use expensive simulation techniques (MCMC) which does not allow its use in ambitious projects like Euclid or SKA. Several challenges will be addressed in this thesis:

- Develop a new DL method to quantify uncertainties, while enjoying theoretical guarantees of coverage. We will rely on conformal quantile regression, a new method derived from theoretical statistics (Romano et al., 2019).

- Generalization: We recently proposed a new architecture of neural networks (the learnets, Ramsi et al., 2020), which has the advantage of including certain properties of the wavelet transform such as exact reconstruction. This type of architecture should provide a solution to the generalization problem.

- The scaling on data of dimension 3 or 4. It will then be a question of extending the results obtained in order to be able to efficiently handle this type of data.

The last challenge of this thesis will be to set up these new tools to solve problems in two large international projects, for dark matter maps with Euclid and SKA.



[1] B. Remy, F. Lanusse, Z. Ramzi, J. Liu, N. Jeffrey and J.-L. Starck, "Probabilistic Mapping of Dark Matter by Neural Score Matching", NeurIPS 2019 Machine Learning and the Physical Sciences Workshop.

[2] Y. Romano E. Patterson E. J. Candès, Conformalized quantile regression. Advances in neural information processing systems 32 NeurIPS, 2019.

[3] Z. Ramzi, JL Starck, T Moreau, P Ciuciu, "Wavelets in the deep learning era", European Signal Processing Conference, accepted submission to the EUSIPCO 2020 conference.
Dark Energy Tomography with the Euclid survey

SL-DRF-21-0206

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire CosmoStat (LCS)

Saclay

Contact :

Valeria Pettorino

Starting date : 01-10-2021

Contact :

Valeria Pettorino
CEA - DRF/IRFU/DAP/LCS


Thesis supervisor :

Valeria Pettorino
CEA - DRF/IRFU/DAP/LCS


Personal web page : https://www.valeriapettorino.com/

Laboratory link : http://www.cosmostat.org/

While the Universe is expanding with increasing velocity, the question of what is causing cosmic acceleration remains unsolved. Acceleration seems to act against gravitational attraction, as if a new source of energy, dubbed dark energy, were responsible for it.

This PhD proposal is meant to contribute to the Euclid mission, to tackle this dilemma by implementing the possibility to test dark energy at different redshifts, or what I refer to here as ‘dark energy tomography’, and integrate it in the Euclid Consortium validated likelihood.

The PhD student will be able to work at the interface between data and theory and concretely collaborate to a large collaboration like the Euclid satellite. Objectives include 1) extending the likelihood software to test dark energy at different redshift epochs, 2) contribute to the collaboration effort on comparing theoretical predictions with data 3) investigate different machine learning methods to reconstruct the dark energy contribution in each redshift bin.

Towards a 3D characterisation of supernova remants in X-rays

SL-DRF-21-0318

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire d’Etudes des Phénomènes Cosmiques de Haute Energie (LEPCHE)

Saclay

Contact :

Fabio Acero

Starting date : 01-10-2021

Contact :

Fabio Acero
CEA - DRF/IRFU/DAP/LEPCHE

0169084705

Thesis supervisor :

Fabio Acero
CEA - DRF/IRFU/DAP/LEPCHE

0169084705

More : http://github.com/facero/sujets2021

X-ray data are multidimensional by nature. For each photon the energy and position is recorded by the X-ray satellite. Here we propose to develop novel techniques to fully exploit the multidimensional nature of the data by combining blind source separation technique with feature learning.
Characterization of SVOM Gamma-Ray Bursts Afterglows using MXT data

SL-DRF-21-0153

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire des spectro-Imageurs spatiaux (LISIS)

Saclay

Contact :

Diego GOTZ

Starting date : 01-10-2021

Contact :

Diego GOTZ
CEA - DRF/IRFU/DAP/LISIS

+33-1-69-08-59-77

Thesis supervisor :

Diego GOTZ
CEA - DRF/IRFU/DAP/LISIS

+33-1-69-08-59-77

More : http://www.svom.fr

SVOM is a mission dedicated to the detection and characterization of Gamma-Ray Bursts (GRBs) and other multi-messenger sources, and it is scheduled for launch in June 2022.



SVOM carries a unique multi-wavelength payload, sensitive from gamma-rays to the visible band, which is complemented on ground by dedicated wide field and narrow field robotic telescopes, distributed over the entire Earth. The SVOM space segment consists of ECLAIRS, a coded mask telescope operating in the 4-150 keV energy range, GRM, a gamma-ray (20 keV-5 MeV) spectrometer, and two follow-up narrow field telescopes, VT (visible) and MXT (0.2-10 keV). The Microchannel X-ray Telescope (MXT) is a compact and light focusing X-ray telescope. The main goal of MXT is to precisely localize the X-ray counterparts of SVOM GRBs and to study in detail their spectral and temporal characteristics.



Gamma-Ray Bursts are issued either by the collapse of a very massive star (> 50 times the mass of the Sun) or by the coalescence of two compact objects (most likely neutron stars). In both scenari a short lived (less than ~100 s) gamma ray emission is followed by a longer lived (hours to days) X-ray emission, that can provide useful information about the GRB environment, the associated emission processes and, possibly, the nature of the GRB progenitors.



The successful PHD candidate will in first place contribute to the analysis the MXT flight model calibration data obtained in summer 2021 at the PANTER X-ray testing facility. In particular, the PHD student will be in charge of producing the MXT spectral response matrices before the SVOM launch, and of updating them during the first two years of the mission, by analyzing the in-flight calibration data.



The PHD student will be part of the MXT science team, act as a SVOM Burst Advocate, and thanks to this experience and to the deep instrumental knowledge acquired she/he will be able to correctly analyze, since the very beginning of the SVOM mission, X-ray afterglow data, and couple them to the multi wavelength data in order to build a clear phenomenological picture of the SVOM GRBs. In fact, the early GRB afterglow phase is still not fully understood, in particular from the point of view of the contribution of the GRB central engine to the so-called “plateau phases” of the afterglow light curve, whose interpretation could lead to a better understanding of the GRB progenitors.

Measuring the growth of massive structures in the distant Universe with deep spectroscopic surveys

SL-DRF-21-0166

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire de Cosmologie et d’Evolution des Galaxies (LCEG)

Saclay

Contact :

Emanuele DADDI

Starting date : 01-10-2021

Contact :

Emanuele DADDI
CEA - DRF/IRFU/DAP/LCEG


Thesis supervisor :

Emanuele DADDI
CEA - DRF/IRFU/DAP/LCEG


Intergalactic magnetic field and gamma ray bursts with CTA

SL-DRF-21-0143

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire d’Etudes des Phénomènes Cosmiques de Haute Energie (LEPCHE)

Saclay

Contact :

Renaud Belmont

Thierry STOLARCZYK

Starting date : 01-09-2021

Contact :

Renaud Belmont
Université de Paris (Paris 7) - DRF/IRFU/DAP/LEPCHE


Thesis supervisor :

Thierry STOLARCZYK
CEA - DRF/IRFU/DAp/LEPCHE

+33 1 69 08 78 12

Personal web page : http://irfu.cea.fr/Pisp/thierry.stolarczyk/

Laboratory link : http://irfu.cea.fr/en/Phocea/Vie_des_labos/Ast/ast_technique.php?id_ast=3709

More : http://www.cta-observatory.org/

The intergalactic magnetic field pervading the cosmic voids is suspected to be a relic field originating from the very first epoch of the cosmic history. The goal of this PhD is to look for signatures of this field in the high-energy data of gamma-ray bursts, and to predict the ability of the future CTA observatory to constrain its properties. This work combines both theoretical modelling and analysis of simulated CTA data.
Multi-physics interaction between exoplanet atmospheres and their host stars

SL-DRF-21-0543

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire de dynamique des étoiles des (Exo) planètes et de leur environnement (LDE3)

Saclay

Contact :

Antonio Garcia Muñoz

Antoine Strugarek

Starting date : 01-10-2021

Contact :

Antonio Garcia Muñoz
CEA - DRF


Thesis supervisor :

Antoine Strugarek
CEA - DRF/IRFU/DAP/LDE3

0169083018

Laboratory link : http://irfu.cea.fr/dap/LDEE/index.php (http://irfu.cea.fr/dap/LDEE/index.php)

The census of known exoplanets includes >4,000 planets in >3,000 systems. The current exoplanet demographics show that exoplanets exhibit a large diversity in their mass, radius (and thus density and composition) and orbital arrangements. The research field moves today from exoplanet detection to the characterization of their atmospheres. The focus of this project is to study numerically the physical mechanisms (3D dynamics, photochemistry, magnetic interactions) that drive the escape of strongly irradiated atmospheres. The project will provide insight into exoplanet systems for which upper atmospheric in-transit observations of e.g. Lyman-alpha, C II, H-alpha, He I at 1083 nm exist but that remain without a proper theoretical interpretation. Our priority is thus to develop physically-motivated models that embrace the complexity of these interactions and help place in context the available observables. We see this project as a first step into the development of a versatile and powerful 3D multi-physics model that can become the international reference for exoplanet-host star interactions.

Cosmology - Clusters of galaxies - Artificial intelligence

SL-DRF-21-0332

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire de Cosmologie et d’Evolution des Galaxies (LCEG)

Saclay

Contact :

Marguerite PIERRE

Starting date : 01-10-2021

Contact :

Marguerite PIERRE
CEA - DRF/IRFU/SAp/LCEG

0169083492

Thesis supervisor :

Marguerite PIERRE
CEA - DRF/IRFU/SAp/LCEG

0169083492

Personal web page : https://sci.esa.int/s/WLg9apw

Laboratory link : http://irfu.cea.fr/dap/Phocea/Vie_des_labos/Ast/ast_groupe.php?id_groupe=972

More : Projet XXL : http://irfu.cea.fr/xxl

Clusters of galaxies are the most massive entities in the universe. As such, they constitute powerful cosmological probes.

The XLL survey is the largest programme of the European satellite XMM (X-ray band). It has enabled the detection of several hundreds of galaxy clusters.

The goal of the PhD is to perform the cosmological analysis of the complete XXL cluster sample by using innovative machine learning techniques.

SL-DRF-21-0661

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire de modélisation des plasmas astrophysiques (LMPA)

Saclay

Contact :

Patrick Hennebelle

Matthias GONZALEZ

Starting date : 01-10-2021

Contact :

Patrick Hennebelle
CEA - DRF/IRFU/DAp

0169089987

Thesis supervisor :

Matthias GONZALEZ
Université de Paris - DRF/IRFU/DAp/LMPA

33 1 69 08 17 79

Personal web page : http://irfu.cea.fr/Pisp/matthias.gonzalez/

Laboratory link : http://irfu.cea.fr/Sap/Phocea/Vie_des_labos/Ast/ast_groupe.php?id_groupe=1250

Study of Polarimetric Bolometer Arrays with Spectroscopic Capabilities for Astrophysics

SL-DRF-21-0652

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire des spectro-Imageurs spatiaux (LSIS)

Saclay

Contact :

Louis RODRIGUEZ

Vincent REVERET

Starting date : 01-10-2021

Contact :

Louis RODRIGUEZ
CEA - DRF/IRFU/DAP/LSIS

0169086948

Thesis supervisor :

Vincent REVERET
CEA - DSM/IRFU/DAp/LSIS

01 69 08 74 02

Laboratory link : http://irfu.cea.fr/dap/Phocea/Vie_des_labos/Ast/ast_groupe.php?id_groupe=2755

The Herschel Space Observatory, launched in 2009, has revolutionized our vision of the "Cold Universe". In particular, it has radically changed our understanding of star formation by highlighting the ubiquitous filamentary structures of gas and dust and their essential role in the very early stages of star formation.

On the other hand, the observations of the Planck satellite (also launched in 2009) in polarimetric mode have highlighted the presence of magnetic fields on large scales in molecular clouds. In these regions, the filaments can be either parallel to the magnetic field (low density filaments) or perpendicular (very dense filaments). But many questions remain.

In order to understand the set of physical processes implemented in these stellar formation zones and to explain the links with the complex structure of the surrounding interstellar medium (ISM), new extremely sensitive instruments must be developed in the submillimeter domain. It seems necessary, on the one hand, to be able to finely characterize the magnetic field (via the detection of polarized light) in several spectral bands and, on the other hand, to detect the presence of several tracers of the ISM via the emission of certain spectral lines (C+ at 158 µm in particular). These observations, made from space or aboard stratospheric balloons, strongly constrain the volume and mass of the on-board charge. The idea of gathering one or more light analysis functions within a compact detector is a step in this direction.

In this context, CEA has been developing for a few years now ultra-sensitive submillimeter bolometer arrays, capable of measuring the polarization within pixels, without the help of external polarizers. Developed in close collaboration with CEA-LETI in the framework of the B-BOP instrument on the SPICA observatory, this technology is in line with the developments of the Herschel-PACS detectors. These bolometers are developed in the framework of Labex Focus, 2 R&T CNES and ESA funding.

In 2019, a thesis defended at the laboratory demonstrated that it was possible to add spectroscopic capacity to these new generation arrays by coupling the detector arrays to a compact Fabry-Perot interferometric system. The experimental demonstration of the complete device remains to be done and this is the core of this thesis topic: to study, implement and characterize experimentally the scientific performances of this compact spectro-imager-polarimeter.

The first step will be to experimentally validate in a cryostat the Fabry-Perot mirror displacement system and to deduce its technical limitations. The second phase will consist in coupling this system to the bolometer arrays in order to produce and characterize the first prototypes of this new type of detectors. Finally, in a third part, the data processing aspect will be studied in order to extract the scientific signal as well as possible and to propose an adequate calibration.

This work may also pave the way to more applied applications in medical imaging or in the field of security controls in the TeraHertz band, as proposed by LETI with its developments of room temperature micro-bolometers.
THE ROLE OF MAGNETIC FIELDS IN STAR-FORMING FILAMENTS: USING THE LARGE OBSERVING PROGRAM 'B-FUN' WITH NIKA2-POL TO PAVE THE WAY FOR THE B-BOP POLARIMETER IN SPACE

SL-DRF-21-0871

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire d’études de la formation des étoiles et du milieu interstellaire

Saclay

Contact :

Philippe ANDRÉ

Starting date : 01-10-2021

Contact :

Philippe ANDRÉ
CEA - DRF/IRFU/DAp/LFEMI

0169089265

Thesis supervisor :

Philippe ANDRÉ
CEA - DRF/IRFU/DAp/LFEMI

0169089265

Laboratory link : http://irfu.cea.fr/dap/Phocea/Vie_des_labos/Ast/ast_technique.php?id_ast=3627

More : https://ipag.osug.fr/nika2//Welcome.html

Imaging surveys of nearby molecular clouds at submillimeter wavelengths with the Herschel and Planck satellites have revolutionized our understanding of the link between the structure of the interstellar medium (ISM) and the early phases of the star formation process. The results of these surveys support a paradigm of star formation in which magnetized filaments play a central role. In particular, Herschel results indicate that most (> 75%) cores/stars form in dense, « supercritical » filaments of ~ 0.1 pc width for which the mass per unit length exceeds the critical line mass of nearly isothermal gas cylinders, and Planck polarization data suggest that the formation and evolution of these filaments is largely controlled by magnetic fields.

The low resolution of Planck polarization observations (5’-10’ at best or 0.2-0.4 pc in the nearest star-forming regions) was however insufficient to probe individual cores along filaments and to understand the role of magnetic fields in filament evolution and fragmentation. This is now becoming possible with new imaging polarimeters such as SOFIA-HAWC+ (18'' resolution at 214 microns) and NIKA2-Pol (12’’ resolution at 1.2mm), the polarimeter that will very soon be available on the new millimeter continuum camera NIKA2 of the IRAM 30m telescope (cf. http://www.iram-institute.org/EN/news-astronomers/2016/118.html and http://ipag.osug.fr/nika2/Welcome.html).

A total of 300 hr of NIKA2 guaranteed time have been set aside for a large polarimetric imaging survey of ~ 10 nearby star-forming filaments with NIKA2-Pol (PI : Ph. André).

The goal of the proposed PhD thesis will be to participate in these polarimetric observations and to interpret the polarization results, based on a detailed comparison with numerical simulations of magnetized cloud collapse and fragmentation obtained with the AMR code RAMSES available in the AIM Theory group.
Modelling magnetar formation

SL-DRF-21-0848

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire de modélisation des plasmas astrophysiques (LMPA)

Saclay

Contact :

Jérôme Guilet

Thierry FOGLIZZO

Starting date : 01-10-2021

Contact :

Jérôme Guilet
CEA - DRF/IRFU

06 38 62 46 30

Thesis supervisor :

Thierry FOGLIZZO
CEA - DRF/IRFU/DAP/LMPA

01 69 08 87 20

Magnetars are neutron stars with the strongest magnetic fields known in the universe. The birth of these objects is one of the most studied scenarios to explain some of the most violent explosions: superluminous supernovae, hypernovae and gamma-ray bursts. Moreover, they emit at least some of the fast radio bursts, as demonstrated in 2020 by the detection of a radio burst from a galactic magnetar. The scientific exploitation of the increasingly abundant data from these various objects requires the development of more predictive models. One of the elements that is still incomplete concerns the prediction of magnetars magnetic field and their evolution. Recently, numerical simulations have succeeded in describing the formation of a magnetic field of comparable intensity to magnetars thanks to amplification mechanisms taking place in the first few seconds after the neutron star is formed. However, most of the observational manifestations of magnetars cited above require the magnetic field to survive on much longer time scales (from a few weeks for superluminous supernovae to thousands of years for soft gamma repeaters). This thesis aims to determine how the turbulent magnetic field generated in the first few seconds will evolve to a stable equilibrium state able to explain these observations.
ORIGIN AND DYNAMICS OF PROTOSTAR CLUSTERS

SL-DRF-21-0870

Research field : Astrophysics
Location :

Direction d’Astrophysique (DAP)

Laboratoire d’études de la formation des étoiles et du milieu interstellaire

Saclay

Contact :

Philippe ANDRÉ

Philippe ANDRE

Starting date : 01-10-2021

Contact :

Philippe ANDRÉ
CEA - DRF/IRFU/DAp/LFEMI

0169089265

Thesis supervisor :

Philippe ANDRE
CEA - DRF/IRFU/DAp/LFEMI

0169089265

Most stars are believed to form in groups or clusters (e.g. [6]), and our Sun itself may well have formed in a star cluster ([1]). Moreover, it appears that massive (> 10 Msun) stars may only form accompanied by a cluster of lower-mass stars. Our present theoretical understanding for the formation of protostars is, however, largely limited to single objects within isolated dense cloud cores (cf. [8]). Current theoretical models do not explain the statistical distribution of stellar masses (the so-called initial mass function - IMF - which seems quasi-universal) and neglect dynamical effects, such as interactions and collisions, which may play a crucial role in embedded young star clusters. To understand the origin of stellar masses and the physical conditions for the birth of the Solar system, it is of paramount importance to characterize the detailed properties of young (proto)star clusters, including their dynamics, in nearby molecular clouds. The latter provide ideal laboratories for star formation studies.

The results of extensive imaging surveys of nearby molecular clouds at submillimeter wavelengths with the Herschel space observatory suggest that the non-uniform spatial distribution or clustering of young stellar objects in star-forming regions is largely inherited from the filamentary structure of the parent clouds (cf. [2], [3], and Fig. 1). The Herschel observations do not provide direct constraints on the dynamics of nearby cluster-forming clouds, however. It is also unclear at which evolutionary stage protostars and young stellar objects decouple from the molecular gas of their parent cloud. These issues can now be addressed thanks to a combined analysis of the astrometric data (e.g. proper motions) from the Gaia satellite, the wide-field submillimeter images of the clouds from Herschel (e.g. Fig. 1), and additional millimeter line observations of the gas kinematics with the IRAM telescopes.

The thesis work will divided into four complementary tasks:

1) Systematic analysis of the protostellar content of the Herschel Gould Belt survey (HGBS) database ([3], http://gouldbelt- herschel.cea.fr/archives). The goal here will be to quantify the clustering of protostars in connection with the filamentary structure of the gas in nearby (d < 500 pc) molecular clouds, in a statistical manner. The Herschel data may be supplemented by higher-resolution observations at 350/450 microns with the ArTéMiS camera on APEX (cf. [5]).

2) Analysis of the Gaia DR2 database (e.g. [7]) to quantify the relative proper motions (in the plane of the sky) and internal dynamics of the young star population in the same clouds.

3) Complementary millimeter observational study of a portion of the same clouds in molecular lines with the IRAM telescopes to constrain the kinematics of the gas and assess the relative motions (projected on the line of sight) of protostars and protostellar dense cores (cf. [4]).

4) Comparison of the observational constraints coming from Herschel, Gaia, and IRAM data with numerical simulations of molecular cloud evolution performed with the adaptive mesh refinement code RAMSES available in the theoretical group of the AIM laboratory.
Deep Learning and gamma spectroscopy: a new signal processing approach for CdTe detectors data analysis

SL-DRF-21-0316

Research field : Instrumentation
Location :

Direction d’Astrophysique (DAP)

Laboratoire des spectro-Imageurs spatiaux (LSIS)

Saclay

Contact :

Olivier LIMOUSIN

Starting date : 01-10-2021

Contact :

Olivier LIMOUSIN
CEA - DRF/IRFU/DAP/LSIS

01 64 50 15 03

Thesis supervisor :

Olivier LIMOUSIN
CEA - DRF/IRFU/DAP/LSIS

01 64 50 15 03

This thesis at the interface between nuclear instrumentation and applied mathematics consists in developing and implementing advanced methods for processing spectral data from CdTe Caliste detectors for high-energy photons. These sensors, resulting from fundamental research in space astrophysics, are the basic building block of the Spid-X gamma camera born from joint technological developments between the CEA and the company 3D PLUS. It aims at characterizing radiative environments in the framework of nuclear surveillance, for the safety of nuclear operations or research facilities, or for the dismantling of installations.

The methods studied will use Deep Learning tools with the objective of analyzing gamma spectra acquired in a complex environment inducing spectral distortions, potentially difficult to interpret with classical algorithms.



For this purpose, the PhD student will carry out the following lines of study:

- The identification of radioelements and the measurement of their proportion in the signal with one or several absorbing and scattering materials between the sources and the detector (methods: Monte-Carlo Geant4 simulations, Bayesian neural networks, confidence robust learning and experimentation).

- Determination of the nature of the material crossed and the thickness crossed (methods: adversarial neural networks (GANs), self-encoding, experimentation).

- The application to coded mask imaging methods. Depending on the results obtained in the two previous axes and the resulting discovery space, the methods may be applied to the theme of coded mask methods for gamma-ray imaging.

 

Retour en haut