Skip to main content

This job has expired

Development of machine learning techniques to describe nuclear reactions

Employer
L2IT, Toulouse
Location
Toulouse, France
Salary
Unspecified
Posting live until
16 Jun 2024

View more

Collisions between atomic nuclei can give rise to a variety of processes, including

quasi-fission, fusion-evaporation, and fusion-fission. Understanding these processes is

important for understanding mechanisms linked to nuclear structure, for predicting reactions

of astrophysical interest, and for the synthesis of super-heavy nuclei.

State-of-the-art theoretical calculations (time-dependent density functional theory) offer a

microscopic description of these reactions and of nuclear dynamics more generally. However,

they require considerable computation time, limiting possible applications. Machine learning

models can overcome these limitations by providing surrogate models (a model fitted to

another model) to perform large-scale calculations. For example, the fast model can be used

to take into account all the initial degrees of freedom (impact parameter, orientation of

deformed nuclei) of nuclear collisions.

This thesis proposes to exploit new machine-learning techniques to create an emulator

of microscopic nuclear dynamics codes, enabling a global description of these reactions. This

innovative approach opens the way to a deeper understanding of nuclear reactions, as well

as to a more complete description of super-heavy element synthesis reactions and long-term

nuclear processes in neutron stars.

Keywords: nuclear reactions, quasi-fission, transfer, fusion, machine-learning

Candidate profile: The candidate must have a Master 2 degree (or be in the process of

obtaining one). The candidate must have a solid background in machine learning techniques,

nuclear physics, quantum physics, and numerical techniques.

Funding: ANR

Application: The candidate should send a CV, a letter of motivation, transcripts with rankings

in M1 and M2, and the contact details of two people with whom he or she has worked.

Get job alerts

Create a job alert and receive personalised job recommendations straight to your inbox.

Create alert