! News: - Avis de Soutenance de Mr. Othmane Laousy : "Deep learning methods for localization, segmentation and robustness in medical imaging."
18 - 22 Mars 2024: Conference Analysis on fracturals and networks, and applications, CIRM, Marseille

Stage M2: "Deep Ritz, Problèmes aux Limites et Apprentissage Semi-Supervisé"


Founded in the early 2000's, MICS (former MAS) is the research laboratory in Mathematics and Computer Science at CentraleSupélec. Research at MICS ic concerned with the analysis and modelling of complex systems and data, whether they come from the industry, life or social sciences, financial markets, information technology or networks.




Research Axes


  •  Biomathematics Mathematical Modelling and Statistical Inference for Biological Systems and Data. Applications to precision medicine, neurosciences, molecular biology, genetics, plant science, agronomy, epidemiology


  • Formal methods in computer science and knowledge modelling Formal design of complex systems; ontologies, decisional and fuzzy logic for image interpretation; unified semantics for structured data.


  • Quantitative Finance : Financial market modelling; market microstructure; high frequency data; econophysics; derivatives.


  • Partial Differential Equations and Scientific Computing : Dynamics of ecosystems; inverse problems; massively parallel computing; GPU computing; irregular and fractal boundaries; nonlinear wave equations; shape optimization, irregular and fractal boundaries, inverses problems.


  • Probabilistic Modelling and Statistics of Stochastic Processes Local regularity of stochastic processes; set-indexed processes; statistical properties of graphs; stochastic differential equations and stochastic partial differential equations.


  • Artificial Intelligence : Deep learning; Representation learning; Information retrieval; Generative Adversarial Networks; Bayesian Learning; Explainability of algorithms


  • Decision modelling : Multicriteria decision making, preference learning, knowledge representation and reasoning, explaining decisions, multiobjective optimization

 Application Domains


Industrial systems (aerospace, construction, energy, transportation); life sciences (medicine, molecular biology, genetics, epidemiology)environment (plants, hydrology, landscapes, acoustics); markets and companies (finance, capital markets, business intelligence); information technology and networks (internet, multimedia, knowledge management).



Victor Bouvier, a PhD student under the supervision of Professor Céline Hudelot and funded by a CIFRE collaboration with Sidetrade, has been the recipient of the best paper in Machine Learning and the best student paper award at the conference ECML/PKDD2020, for the paper "Robust Domain Adaptation: Representation, Weights and Inductive Bias", V. Bouvier, P. Very, C. Chastagnol, M. Tami, C. Hudelot.

Véronique Letort was granted by ITMO-Cancer in the call "Apports à l'oncologie des mathématiques et de l'informatique" for herproject RadioPrediTool. In collaboration with Gustave Roussy, the objective is to predict iatrogenic effects after the treatment of pediatric cancers.

MICS is in charge of coodinating the scientific actions for CentraleSupélec within PRISM, the Centre for Precision Medecine in Oncoloy, a collaboration between Gustave Roussy, CentraleSupélec, Paris-Saclay and INSERM, labelled by ANR (French National Research Angency).

Academic partners


 Institut Gustave Roussy, CEA, INRA, INRIA, INSERM, AgroParisTech Cambridge, Oxford, Georg-August-Universität Göttingen, TU München, Sapienza University of Rome, Polytechnic University of Turin, RUDN University, Bar Ilan, University of Tokyo, Doshisha University (Japan), Beihang University, (China), Providence University (Taiwan), University of Washington, University of Michigan, Temple University, Berkeley Lab (USA), University of Connecticut and Bielefeld University.

Industrial partners