Missions

The post-doctoral fellow will work in the software development team of the IJCLab laboratory, together with local representatives of the ATLAS LHC experiment.
He will develop an advanced generative model algorithm for calorimeter simulation (leveraging on existing literature and Proof-of-Concepts). The work will be supervised by David Rousseau.
In the scope of the Large Hadron Collider and future ILC and FCC colliders, the resource usage of detector simulation, in particular calorimeter simulation, has been identified as a bottleneck. This delicate task is traditionally done with Geant4, which tracks the thousands of secondary particles across a finely described virtual detector for each incoming particle. The HEP community has embarked on emulating Geant4 simulation with generative models, initially GAN and VAE, more recently diffusion model or dedicated hybrid architectures. The issue of the simulation of tails over several orders of magnitude, and rare of events, has been paid little attention so far. GenCaloSimExtreme projects aims to focus on the tail description and propose new model HPO, or a post-processing fine-tuning stage. This is mandatory for the deployment of these models for physics analysis.
Current state-of-the-art architectures and open datasets will be reviewed, with a specific focus on tails/rare events description. Particular attention will be paid to in-depth validation of the algorithm with physics-informed features.

Activités

The post-doctoral fellow will be in charge of the following specific activities :
• Develop innovative simulation algorithms for calorimeter in High Energy Physics context, with a particular focus on tails/rare events description
• Communicate on this work via written reports and oral presentations in the HEP community via papers and conference presentations
The expected outcome is a fully functional generative algorithm, with improved description of the tails/rare events with respect to the state of the art

Compétences

Applicants must hold a PhD in particle physics or computer science.
Applicant must have previous experiment in detector simulation or generative models applied to scientific modeling. Experience with calorimeter simulation with generative model would be a plus.
Applicants must also demonstrate teamwork skills and an ability to fit in a multidisciplinary scientific environment. In particular, they must communicate their results to peers while following the rigorous thinking and ethical standards expected in a large-scale scientific project. Any concrete proof of such skills, including previous peer-reviewed publications and conference talks, will be greatly appreciated.
Finally, given the international nature of this project and particle physics at large, a good mastery of written and oral English will be required. French is not necessary.

Contexte de travail

The Irène Joliot-Curie Physics Laboratory of 2 Infinities (IJCLab) is laboratory (UMR) under the supervision of the CNRS (IN2P3), the University of Paris-Saclay and the University of Paris is located on the campus of the University of Paris -Saclay in Orsay. The laboratory is located on the campus of Université Paris-Saclay in Orsay. The campus is located 20 km south of Paris and accessible by RER in 35 minutes.
IJCLab was born in 2020 from the merger of five units (CSNSM, IMNC, IPN, LAL, LPT). The staff comprises nearly 560 permanent (340 engineers, technicians and administrators and 220 researchers and teacher-researchers) and approximately 200 non-permanent, including 120 doctoral students. The research themes of the laboratory are nuclear physics, high energy physics, theoretical physics, astroparticles, astrophysics and cosmology, particle accelerators, energy and the environment and health.
IJCLab has very significant technical capacities (around 280 IT) in all the major fields required to design, develop/implement the experimental devices necessary for its scientific activity, as well as the design, development and use of instruments.
For this specific project, IJClab hosted key authors of the recently published first GAN-based fast simulation of the ATLAS electromagnetic calorimeter ( CSBS 8, 7 (2024)) .
IJCLAB bases its recruitment policy on the promotion of equality, diversity and inclusion. Essential values, they allow the professional development of agents, who are real actors in a collective success, but also the development of the laboratory itself.
The successful candidate will have access to IJCLab computing resources, to on-campus resources (Ruche, Jean Zay) as well as to IN2P3 Compute Center in Lyon (an LCG Tier1).
Depending of his background, he will may be registered in ATLAS and thus will have access to CERN computing resources and all ATLAS worldwide compute powers.

Le poste se situe dans un secteur relevant de la protection du potentiel scientifique et technique (PPST), et nécessite donc, conformément à la réglementation, que votre arrivée soit autorisée par l'autorité compétente du MESR.