← Retour aux offres

Thèse CEMEF 2020 : "Physics-Informed Machine Learning in the context of seismic imaging."

Postée le 18 mai

Lieu : Location: mainly Centre de Géosciences, Fontainebleau, France. · Contrat : CDD · Rémunération : GROSS ANNUAL SALARY : about 27k€ €

Société : MINES ParisTech – CEMEF

Leader mondial dans le domaine des matériaux, des procédés et de leur modélisation, le CEMEF réalise une recherche partenariale avec l'industrie et forme des experts (doctorats et mastère spécialisé).
Pluridisciplinaire, le CEMEF étudie tous les types de matériaux (métaux et alliages métalliques, polymères synthétiques et issus de la biomasse, composites, béton...) et de procédés industriels en s’intéressant aux différentes étapes de la vie du matériau :
élaboration -> mise en forme -> traitements thermiques -> assemblage -> propriétés d'usage -> recyclage
Elément majeur d’innovation, sa compréhension du comportement des matériaux aux différentes échelles, de l'interaction outil-matière, de l’impact des procédés sur les propriétés finales permet l’optimisation et le développement de procédés de fabrication de plus en plus complexes et contraints (technologie, énergie, environnement…).
Il s’appuie sur des approches couplant techniques expérimentales et modélisations numériques
Le CEMEF, c'est plus de 160 ingénieurs, chercheurs, étudiants, techniciens, animés par la même curiosité face à la complexité des phénomènes, la même envie de se dépasser pour agrandir le champ des connaissances, le même goût du défi scientifique.
C’est un centre de recherche de MINES ParisTech, associé au CNRS.
Spécialisations : polymères, bioplastiques, composites, métaux, alliages métalliques, transformation des matériaux, physique, mécanique, thermique, propriétés d'emploi, tenue en service, modélisation numérique, développement logiciel, calcul intensif, surfaces.

Description du poste

OVERVIEW

Within the MINDS project (Mines Initiative for Numerics and Data Science) developed at MINES ParisTech (Paris School of Mines), the objective is to fill the gap between Machine Learning and Physics-based approaches.

Machine Learning is growing very rapidly. After a possible learning step, the objective is to let the data speak. These approaches tend to forget the more traditional physics-based approaches. The objective of the PhD thesis is to develop, in the context of seismic imaging, an intermediate approach to preserve the physics [1]. Currently, the main contributions of Machine Learning to seismic processing are related to pre-processing steps (de-noising, picking, ...) but not really yet to the imaging part (determining the Earth's properties from surface measurements, a highly non-linear problem). The explicit introduction of physics within Machine Learning should fill this gap.

DETAILS

In 2019, Raissi et al., demonstrated how it is possible to combine Machine Learning approaches with more traditional physics approaches (Physics-Informed Neural Networks, PINN) [3]. The applications are related to the resolution of partial differential equations (i.e. direct problems) as well as to the resolution of inverse problems (determining the main parameters controlling the physical phenomena, for example the wave propagation, from a set of observations). The later approach will be developed here.
On the one hand, deep neural networks are able in theory to describe any functions. Learning is usually a complex task and in physics-related problems, observations are rare and expensive to acquire. On the other hand, Machine Learning does not usually consider physics- based equations, a very useful source of information. As proposed in [3], a modified loss function in the neural networks contains several terms to ensure that the data predict the observations and that the laws of physics are fulfilled. This second term can be seen as a regularisation term, essential in practice to avoid any over-fitting in the case of noisy data. The auto-differentiation (back-propagation of the errors) within the neural networks provides a way to estimate the optimal parameters.

Plus d’infos sur :
https://www.cemef.mines-paristech.fr/en/open-phd-position-physics-informed-machine-learning-in-the-context-of-seismic-imaging/

Profil recherché

The candidate should have a strong background in maths and physics. He/she should have a clear interest for Machine Learning and for geophysical applications, in particular in the context of seismic imaging. He/she should have a strong experience in scientific programming. It is appreciated if he/she also has some knowledge on high performance computing (HPC). It is essential to be fluent in English speaking and writing.

Pour postuler :

Candidatures en ligne uniquement sur :
http://www.recruitment.cemef.mines-paristech.fr/phd