ANR JCJC GAME – GAussian process modeling of MEchanical random fields: a complete study from simulation to identification

This 4-years project (starting in November 2023) is funded by the French National Research Agency (ANR, abbreviation in French), as part of the call for young researcher projects.

Consortium

• Université Polytechnique Hauts-de-France (UPHF)
  • A. F. López-Lopera (PI), CERAMATHS
  • F. Massa, LAMIH
  • L. Reding, CERAMATHS
  • I. Turpin, CERAMATHS
• Mines Saint-Etienne (EMSE):
  • J. Bruchon, SMS
  • R. Le Riche, LIMOS, CNRS

Summary

The ANR JCJC GAME project focuses on simulation and identification problems in mechanical systems. Our main interest are:

• Encoding expert knowledge into kernels modeling mechanical fields.
• Systems where inputs are fields (or distributions), and outputs are time-series.
• Theoretical guarantees of estimators for Gaussian processes with functional inputs.
• Parameter identification associated to mechanical fields from measurable quantities.

The project proposes both innovative theoretical and numerical contributions that are of interest in research fields such as applied mathematics, machine learning and mechanics (but not limited to them). Our theoretical contributions are based on Gaussian process models with (multivariate or distribution) functional inputs and functional outputs. Since they may be independent of the mechanical applications, the resulting models can then be used in other research fields.

Keywords. Mechanical codes, stochastic modeling and simulation, uncertainty quantification, functional data, kernel design, machine learning, reverse-engineering.

Funding opportunities

• 1 Master internship (6 months, starting in Spring 2024)
  • Status: already allocated
• 1 PhD thesis (3 years, starting in Fall 2024 or as soon as possible)
  • Status: already allocated

Both offers can be found [here]

To acknowledge this project: “This work was supported by the project GAME (ANR-23-CE46-0007) of the French National Research Agency (ANR).”