En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)
Portail > Offres > Offre UMR6534-AURGON-018 - CDD chercheur (H/F) : Algorithmes d'Apprentissage en Physique Fondamentale et Appliquée

Postdoctoral researcher (M/F) : Machine Learning in fundamental and applied physics

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

Date Limite Candidature : mercredi 12 octobre 2022

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 : UMR6534-AURGON-018
Workplace : CAEN
Date of publication : Wednesday, September 21, 2022
Type of Contract : FTC Scientist
Contract Period : 36 months
Expected date of employment : 1 November 2022
Proportion of work : Full time
Remuneration : From 4 353,15 € gross per month depending on experience
Desired level of education : PhD
Experience required : Over 10 years


The objective of this project is to develop a new research and engineering axis around artificial intelligence (AI) techniques at LPC CAEN. It proposes to advance along two axes : one is to strengthen the current research in fundamental physics and the other, to develop novel methods and applications in the nuclear industry. More specifically, the applicant wiil have to test and deploy new AI methods to improve our understanding of the fundamental properties of the neutrino with the KM3NeT/ORCA experimental data, and to develop similar techniques towards real-time dosimetry.

Despite enormous experimental progress, the neutrino is still the most enigmatic particle of the Standard Model of Particle Physics. After the successful establishment of the three neutrino flavor mixing, the next decade could be as promising with a worldwide experimental effort to construct several "mega-detectors" in order to reach the next level of precision. In this context, KM3NeT (Cubic Kilometer Neutrino Telescope) occupies a unique position : This neutrino telescope installed in the Mediterranean Sea at a depth of more than 2000m, is dedicated to multi-messenger astronomy of high energy neutrinos and for the study of atmospheric neutrino oscillations. It comprises an array of several hundred detection units to detect the Cherenkov light originating from neutrino interactions. With a start of the exploitation phase in final configuration scheduled for 2025, KM3NeT/ORCA will be able to obtain a first indication of the mass hierarchy in only 3 years. As part of the KM3NeT collaboration, the LPC-CAEN contributes to the building and deploying of detector lines in the Mediterranean Sea. It must also develop the necessary analysis tools to process the vast quantities of data that it is accumulating since early 2020. This project intends therefore to develop and exploit novel AI techniques in order to improve the current analyses and strengthen KM3NeT's competitiveness on measuring neutrino oscillation parameters.

The reliability of radiation measurements is a central aspect in many activities of the nuclear and medical sectors. In the case of neutrons, many of the survey meters used are not very accurate (due to scattering effects, size of the systems and complexity of the sources) which often leads to conservatism in dose calculations. The project proposes the development of AI techniques around a new type of multi-mode detector which allows to sample neutron and gamma fluxes by segmenting the active volume of the system into voxels. This system captures at the same time the direction of emission, the energy, the spatial distribution of the radiations and the possibility to precisely reconstruct the effective dose for directional and isotropic particle fluxes.

Finally, it aims to significantly increase the range of competences and the impact of the LPC CAEN in the current scientific projects, while developing applications with the actors of the Normandy region. It is also intended to promote teaching and training at the master and doctoral levels that will give new generations of students the opportunity to have access to training through research based on transferable and forward-looking cutting-edge tools.


For the KM3NeT project, the selected candidate will focus on the following aspects of machine learning techniques:
- Development of a framework for the design, deployment and maintenance of the
algorithms used in the KM3NeT analyses
- Design of new distribution approximation algorithms (such as GANs, auto-encoders or Normalising Flow) to increase sampling by avoiding some of the costly steps in the simulation.
- Exploration of self-learning and pruning techniques for model optimization.

For dosimetry activities, the selected candidate will focus on:
- Optimization and maintenance of algorithms for dosimetry using the same framework as for KM3NeT.
- Optimization of the number of measurements and counting time by simulating virtual campaigns and using artificial intelligence agents to optimize the problem. The objective is to progress towards real-time dosimetry.
- Collaboration with local SMEs to set up demonstrations, using their know-how in virtual reality.

In addition, the candidate will have to supervise two PhD thesis during the project : One thesis on KM3Net data analysis and one thesis on neutron dosimetry. He(She) will also have to supervise M1 or M2 internships (at least 1 per year).


Candidates are expected to have a PhD thesis in nuclear physics and/or experimental particle physics. Recognized skills in Numerical method and machine learning are required.

Work Context

The successful candidate will be assigned to the KM3NeT team of the GrAMM group, within the Laboratoire de Physique Corpusculaire de Caen. This team is composed of 4 permanent members: two lecturers and two CNRS research. Its research activities concern the simulation and analysis of data from the KM3NeT telescope. A thesis is in progress on the study of atmospheric neutrino oscillations with KM3NeT/ORCA data. In addition, the Medical and Industrial Applications group of LPC Caen has been developing instruments dedicated to radiation protection for many years. In particular, the DOENUT project aims to develop a prototype transportable neutron spectrometer/debitometer for the nuclear power industry. The decompartmentalized functioning of the LPCC between the different research groups and services is a favorable environment for the creation of a synergy around the development of the new neutron/gamma dosimeter/imager.

The LPC CAEN, with about 87 staff, is a joint research unit (UMR 6534) depending on three supervisory bodies: CNRS, the University of Caen Normandy (UCN) and the National Engineering School of Caen (ENSICAEN). It is located on the Campus 2 of the University of Caen (Campus Côte de Nacre) and is part of the research park of ENSICAEN (www.lpc-caen.in2p3.fr/).

Constraints and risks

Work under ionizing radiation,
Travels in France and abroad.

Additional Information

Contract financed by the Normandy Region.

We talk about it on Twitter!