Nested Metasampling for Material Science

Nested Metasampling for Material Science

Directeur(s) ou directrice(s) : Martino Trassinelli, martino.trassinelli@insp.jussieu.fr; Fabio Finocchi, fabio@insp.jussieu.fr
Financement : Sorbonne Center of Artificial Intelligence (Sorbonne Université)
Description : https://w3.insp.upmc.fr/wp-content/uploads/2025/04/PhD_QNEST_2025.pdf
Début : 2025
Fin :
Doctorant.e :
Equipe(s) : Agrégats et surfaces sous excitations intensesOxydes en basses dimensions
Page des thèses de(s) l'équipe(s) : Agrégats et surfaces sous excitations intensesOxydes en basses dimensions
Etat de la thèse : Thèse et Thèse proposée

 

We propose an interdisciplinary project involving different fields such as Bayesian data analysis, statistical physics and applied mathematics for the efficient exploration and minimization of complex scalar functions. The project is centered on the Nested Sampling optimization algorithm based on Bayesian statistics to explore the relevant parameter space recursively effectively, obtain the topology thereof, and assess the quality and reliability of the obtained optimum within a pre-determined statistical error.

Detailed description here.