En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)
Portail > Offres > Offre UMR7315-ALEBOU-003 - Chercheur (H/F) en science des matériaux et machine learning

Post-doctoral position (F/M) in materials science and machine learning

This offer is available in the following languages:
Français - Anglais

Date Limite Candidature : vendredi 3 février 2023

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

General information

Reference : UMR7315-ALEBOU-003
Nombre de Postes : 1
Workplace : LIMOGES
Date of publication : Friday, January 13, 2023
Type of Contract : FTC Scientist
Contract Period : 13 months
Expected date of employment : 1 April 2023
Proportion of work : Full time
Remuneration : 2805,35 € monthly gross salary
Desired level of education : PhD
Experience required : Indifferent


Deep-learning analysis of X-ray diffraction data : application to spatial strain profile retrieval in irradiated materials


In the recent years,artificial intelligence, and more specifically, deep neural networks have thrived and grew out of the purely computer-science domain and infiltrated almost all scientific disciplines. The present project deals with the case of the inversion of X-ray diffraction (XRD) data so as to determine real space nanostructural characteristics, from reciprocal space data (i.e., XRD). This task is well know to be hampered by the “phase problem”, which describes the fact that the phase of the diffracted x-rays is lost in the diffraction experiment, so that real space information can not be retrieved from a simple inversion of the XRD data. This issue is usually circumvented by fitting a parameterized physical model to the data.

While such approaches have proven very efficient, their main drawback is the time needed to reach convergence, which is of the order of a few minutes, up to several hours depending on the complexity of the problem. Moreover, the simulation has to be conducted by an expert to avoid the algorithm being trapped in secondary minima during the optimization. These characteristics become unbearable when dealing with large amounts of data which are now easily acquired with laboratory diffractometers and, even more critically, at synchrotron facilities.
In the present project we shall focus on the determination of strain profiles in ion-irradiated materials. This problematic is critical in several domains, ranging from the implantation of semiconductors, to nuclear materials, encompassing spatial applications. Recently, a proof of concept has been published that demonstrates the feasibility of the approach [https://iopscience.iop.org/article/10.1088/2632-2153/acab4c]. The work will consist in (i) generating a numerical XRD data base representative of actual data, (ii) designing a convolutional neural network (CNN) and (iii) training and optimizing the CNN using the data generated. The performances of the CNN will be benchmarked against simulation performed on real data that we collected during the last 10-15 years.

The computer codes will be developed using the Python programming language, using the NumPy and SciPy libraries for scientific computing, and the TensorFlow library with the Keras front end, for the implementation of the neural networks. All developments will be implemented on dedicated workstations available at IRCER.


The applicant should hold a Ph.D. in physics or materials sciences and be familiar with X-ray diffraction or crystallography, and with Python programming. Knowledge in machine learning/deep learning would be appreciated but is not mandatory.
Important: the applicant should hold a PhD since less than 2 years.

Work Context

Location: CNRS laboratory IRCER, Limoges, France

Additional Information

Less than 2 years work experience.

We talk about it on Twitter!