AI for Science: Physics Based Deep Learning for Modeling Complex Dynamics. Application to Climate

10/01/2022 – 11/01/2022 all-day

Offre en lien avec l’Action/le Réseau : – — –/– — –

Laboratoire/Entreprise : Sorbonne Universite – Equipe Machine Learning and
Durée : 6 mois
Contact :
Date limite de publication : 2022-01-10

Contexte :
AI for science is concerned with the exploration of machine learning for scientific computing in domains traditionally dominated by physics models (first principles). We consider here the modeling of complex dynamical systems characterizing natural phenomena, a recent and fast growing research direction with a focus on climate modeling applications and with the objective of combining model based physics and machine learning approaches.

Sujet :

The global objective is the development of new models integrating physics prior knowledge and deep learning (DL) for the modeling of spatio-temporal dynamics characterizing physical phenomena such as those underlying earth-science and climate observations. The classical modeling tools for such dynamics in physics and applied mathematics rely on partial differential equations (PDE). We then consider situations where the physical prior background is provided by PDEs. Two main directions will be explored:

– Hybrid systems – Interfacing Deep neural Networks (DNNs) and PDE

– Domain generalization for deep learning as dynamical models
Depending on the progress on the first topic, one will consider the issue of domain generalization of hybrid models.

– Application to climate data: the application will target the modeling of the dynamics of ocean circulation, which is a component of climate models.

Profil du candidat :
Master or engineering degree in computer science or applied mathematics.

Formation et compétences requises :
The candidate should have a strong scientific background with good technical skills in programming.

Adresse d’emploi :
Machine Learning and Information Access team – MLIA –, Sorbonne University, 75005 Paris, Fr

Document attaché : 202112031804_2021-12-MLIA-Internship-Deep-Learning-Physics.pdf