SCALES

Statistical ChAracterization of multi-scaLE complex Systems with information theory
Projet ANR
Approval no ANR-21-CE46-0011-01
Start: 2022
End: 2025

Statistical physics shows strong benefits when describing multi-scale complex systems such as: fluid turbulence, climate or neural signals.  In particular, Information Theory exhibits strong potentialities in the study of complex systems due to its power to characterize non-linear behaviors. Moreover in the last years, AI models have been strongly developed to deal with a large number of scientific questions, and more particularly complex systems. Thus, SCALES proposes to combine this IT framework with AI models to characterize interactions among the scales of complex systems.

Context

In nature, a large number of systems and processes present non-linear and multi-scale structures and behaviors, and are thus considered as complex systems. Because of these properties, models aiming at describing or predicting the behavior of complex systems need to take into account high-order statistics in a whole range of scales.

In this project, I propose to develop a new statistical description of multi-scale coupling and interactions based on Information Theory (IT), which can be directly linked to complexity and then to the structure and state of this kind of systems. I also propose the combination of this statistical characterization with Artificial Intelligence (AI) in order to develop new multi-scale learning-based approaches for modeling complex systems. As case study, I focus on the characterization and modeling of ocean surface dynamics, since it is a major issue in oceanographic and climate research where both non-linearity and multiple scales are present.

 

Objectives

The SCALES project presents three main goals. First, the development of a statistical description of multi-scale couplings and interactions based on Information Theory. Second, the formulation of multi-scale Deep Learning (DL) models based on the Information Theory framework previously developed. Finally, the third goal of SCALES is the combination of IT and DL for studying fluid turbulence and oceanic flows.

 

 

Expected results

The main expected results are:

  1. The development of a multi-scale Information Theory framework for multivariate anisotropic and inhomogeneous images.
  2. The formulation of a framework for measuring causality interactions between scales in time-series of 2D fields.
  3. The identification of the most adapted combination of multiscale IT metrics with DL models for probing multi-scale high order statistics.
  4. The conception of a new multi-scale and high-order statistics DL model with multi-scale metrics and multi-scale architecture.
  5. A complete statistical description of ocean Lagrangian dynamics from different datasets, and their comparison.
  6. The development of a new DL model for ocean studies with three main novelties: multi-scale, high-order statistics, physically sound

 

Next step

One researcher from Ifremer/LOPS will contribute to SCALES: B. Chapron, currently co-PI of the ERC Synergies STUOD, is a first-line international expert in the field of physical oceanography. 

Anr

Contacts

- Carlos Granero Belinchon (PI) will provide its expertise in signal processing, the characterization of multi-scale and non-linear behaviors, Information Theory and self-similarity and turbulence.

- Ronan Fablet, PI of the current ANR chair OceaniX, will provide its expertise in data science and deep learning for oceanography applications.

Project challenges
/en/research-innovation/issues/digital
Digital transition
/en/research-innovation/issues/environmental
Environmental transition
/en/research-innovation/issues/fundamental-research-transition
Fundamental research transition
Sustainable Development Goals
ODD 13 : Lutte contre les changements climatiques et leurs répercussions
SDG 13: Climate Action
The project website
Site du projet

Project news

Il n'y a pas d'actualité en ce moment.