Christian SOIZE annuaire Université Gustave Eiffel

Christian Soize, Professor Emeritus at University Gustave Eiffel, performs his research in probability theory and mathematical statistics. After earning his PhD from the University “Pierre et Marie Curie” in 1979, he began as a researcher at ONERA and, in 2001, became Professor at the University of Marne-la Vallée (now University Gustave Eiffel).
He has been recognized with several awards, including a prize from the French Academy of Sciences (1985), the IACM Award Computational Mechanics (2018), and the Alfred M. Freudenthal Medal from ASCE (2022). Christian Soize has delivered numerous lectures and contributed extensively through publications and conference presentations.
His research is in the field of applied mathematics, at the interface of probability theory and mathematical statistics, stochastic modeling, and numerical simulation for engineering sciences, particularly theoretical and computational mechanics. It has focused on the mathematical modeling of random phenomena and stochastic dynamical systems, the theory of stochastic fields, and the probabilistic modeling of uncertainties in complex systems, including multiscale modeling, uncertainty quantification, and statistical inverse problems. In particular, a nonparametric probabilistic approach to uncertainty for modeling random operators, based on random matrix theory, has been developed. More recently, the research has addressed probabilistic and statistical learning, as well as machine learning, for high-dimensional problems with limited datasets, while also exploring probabilistic learning formulations adapted to quantum computer.

Current research fields are:
Statistical learning, Probabilistic learning on manifolds, Machine learning, Optimization under uncertainties, Quantum computer formulation.
Uncertainty quantification, stochastic modeling of uncertainties in computational sciences and engineering, their propagation and their quantification solving stochastic inverse problems.
Stochastic multi-scale modeling and application to microstructures of heterogeneous materials.
Computational science, computational mechanics, linear and nonlinear structural dynamics, structural acoustics, vibroacoustics and coupled systems.
Computational stochastic dynamics for linear and nonlinear dynamical systems.

Statistical Learning, Probabilistic Learning, Machine Learning, and Optimization under uncertainties.
Uncertainty Quantification, stochastic modeling of uncertainties in computational sciences and engineering, their propagation and their quantification solving stochastic inverse problems.
Stochastic multi-scale modeling and application to microstructures of heterogeneous materials.
Computational mechanics, linear and nonlinear structural dynamics, structural acoustics, vibroacoustics and coupled systems.
Computational stochastic dynamics for linear and nonlinear dynamical systems.

Mes dernières références

My latest references

Articles, livres et chapitres de livre

Articles, books, and book chapters

PLoM_Matlab_V1_2024

Softwares PLoM (Probabilistic Learning on Manifolds) in Matlab

https://github.com/ChristianSoize/PLoM_Matlab_V1_2024

The PLoM (Probabilistic Learning on Manifolds) software is a MATLAB-based toolbox designed to perform advanced probabilistic modeling and learning on manifolds. It is particularly useful for analyzing high-dimensional data from small datasets, reducing dimensionality, and capturing complex relationships in data through a probabilistic approach. The toolbox provides various types of analyses that facilitate the construction of probabilistic models on manifolds and statistical surrogate models, making it a powerful tool for researchers and practitioners in fields such as scientific machine learning, data science, and applied mathematics.

The development of the PLoM software is grounded in the research published in the articles listed at the end of this file. Numerous applications of this software can also be found in the published literature.

The software's capabilities are demonstrated through 9 Benchmarks. Each Benchmark illustrates a specific type of analysis that can be performed, with fully defined data sets and the associated results generated by PLoM.

PLoM_Matlab_V1_2024

Softwares PLoM (Probabilistic Learning on Manifolds) in Matlab

https://github.com/ChristianSoize/PLoM_Matlab_V1_2024

The PLoM (Probabilistic Learning on Manifolds) software is a MATLAB-based toolbox designed to perform advanced probabilistic modeling and learning on manifolds. It is particularly useful for analyzing high-dimensional data from small datasets, reducing dimensionality, and capturing complex relationships in data through a probabilistic approach. The toolbox provides various types of analyses that facilitate the construction of probabilistic models on manifolds and statistical surrogate models, making it a powerful tool for researchers and practitioners in fields such as scientific machine learning, data science, and applied mathematics.

The development of the PLoM software is grounded in the research published in the articles listed at the end of this file. Numerous applications of this software can also be found in the published literature.

The software's capabilities are demonstrated through 9 Benchmarks. Each Benchmark illustrates a specific type of analysis that can be performed, with fully defined data sets and the associated results generated by PLoM.

PLoM_Python_V1_2024

Softwares PLoM (Probabilistic Learning on Manifolds) in Python
https://github.com/ChristianSoize/PLoM_Python_V1_2024

The "PLoM_Python_V1_2024" package is the Python conversion of the Matlab code "PLoM_Matlab_V1_2024," which is available on GitHub.

The PLoM (Probabilistic Learning on Manifolds) software is a Python-based toolbox designed to perform advanced probabilistic modeling and learning on manifolds. It is particularly useful for analyzing high-dimensional data from small datasets, reducing dimensionality, and capturing complex relationships in data through a probabilistic approach. The toolbox provides various types of analyses that facilitate the construction of probabilistic models on manifolds and statistical surrogate models, making it a powerful tool for researchers and practitioners in fields such as scientific machine learning, data science, and applied mathematics.

The development of the PLoM software is grounded in the research published in the articles listed at the end of this file. Numerous applications of this software can also be found in the published literature.

The software's capabilities are demonstrated through 9 Benchmarks. Each Benchmark illustrates a specific type of analysis that can be performed, with fully defined data sets and the associated results generated by PLoM.

PLoM_Python_V1_2024

Softwares PLoM (Probabilistic Learning on Manifolds) in Python
https://github.com/ChristianSoize/PLoM_Python_V1_2024

The "PLoM_Python_V1_2024" package is the Python conversion of the Matlab code "PLoM_Matlab_V1_2024," which is available on GitHub.

The PLoM (Probabilistic Learning on Manifolds) software is a Python-based toolbox designed to perform advanced probabilistic modeling and learning on manifolds. It is particularly useful for analyzing high-dimensional data from small datasets, reducing dimensionality, and capturing complex relationships in data through a probabilistic approach. The toolbox provides various types of analyses that facilitate the construction of probabilistic models on manifolds and statistical surrogate models, making it a powerful tool for researchers and practitioners in fields such as scientific machine learning, data science, and applied mathematics.

The development of the PLoM software is grounded in the research published in the articles listed at the end of this file. Numerous applications of this software can also be found in the published literature.

The software's capabilities are demonstrated through 9 Benchmarks. Each Benchmark illustrates a specific type of analysis that can be performed, with fully defined data sets and the associated results generated by PLoM.