Search Results for author: Ronan Fablet

Found 44 papers, 19 papers with code

Online Calibration of Deep Learning Sub-Models for Hybrid Numerical Modeling Systems

no code implementations17 Nov 2023 Said Ouala, Bertrand Chapron, Fabrice Collard, Lucile Gaultier, Ronan Fablet

Most of these efforts in defining hybrid dynamical systems follow {offline} learning strategies in which the neural parameterization (called here sub-model) is trained to output an ideal correction.

Neural SPDE solver for uncertainty quantification in high-dimensional space-time dynamics

no code implementations3 Nov 2023 Maxime Beauchamp, Ronan Fablet, Hugo Georgenthum

Recent advancements in deep learning also addressed this issue by incorporating data assimilation into neural architectures: it treats the reconstruction task as a joint learning problem involving both prior model and solver as neural networks.

Gaussian Processes

Gradient-free online learning of subgrid-scale dynamics with neural emulators

no code implementations30 Oct 2023 Hugo Frezat, Ronan Fablet, Guillaume Balarac, Julien Le Sommer

In this paper, we propose a generic algorithm to train machine learning-based subgrid parametrizations online, i. e., with $\textit{a posteriori}$ loss functions for non-differentiable numerical solvers.

Training neural mapping schemes for satellite altimetry with simulation data

no code implementations19 Sep 2023 Quentin Febvre, Julien Le Sommer, Clément Ubelmann, Ronan Fablet

Here, we leverage both simulations of ocean dynamics and satellite altimeters to train simulation-based neural mapping schemes for the sea surface height and demonstrate their performance for real altimetry datasets.


A VAE Approach to Sample Multivariate Extremes

1 code implementation19 Jun 2023 Nicolas Lafon, Philippe Naveau, Ronan Fablet

We illustrate the relevance of our approach on a synthetic data set and on a real data set of discharge measurements along the Danube river network.

Reduction of rain-induced errors for wind speed estimation on SAR observations using convolutional neural networks

no code implementations16 Mar 2023 Aurélien Colin, Pierre Tandeo, Charles Peureux, Romain Husson, Ronan Fablet

By carefully building a large dataset of SAR observations from the Copernicus Sentinel-1 mission, collocated with both GMF and atmospheric model wind speeds as well as rainfall estimates, we were able to train a wind speed estimator with reduced errors under rain.

Scale-aware neural calibration for wide swath altimetry observations

no code implementations9 Feb 2023 Quentin Febvre, Clément Ubelmann, Julien Le Sommer, Ronan Fablet

Sea surface height (SSH) is a key geophysical parameter for monitoring and studying meso-scale surface ocean dynamics.

Inversion of sea surface currents from satellite-derived SST-SSH synergies with 4DVarNets

no code implementations23 Nov 2022 Ronan Fablet, Bertrand Chapron, Julien Le Sommer, Florian Sévellec

This is however limited to the surface-constrained geostrophic component of sea surface velocities.

Neural Fields for Fast and Scalable Interpolation of Geophysical Ocean Variables

1 code implementation18 Nov 2022 J. Emmanuel Johnson, Redouane Lguensat, Ronan Fablet, Emmanuel Cosme, Julien Le Sommer

Optimal Interpolation (OI) is a widely used, highly trusted algorithm for interpolation and reconstruction problems in geosciences.

Deep learning for Lagrangian drift simulation at the sea surface

no code implementations17 Nov 2022 Daria Botvynko, Carlos Granero-Belinchon, Simon Van Gennip, Abdesslam Benzinou, Ronan Fablet

We address Lagrangian drift simulation in geophysical dynamics and explore deep learning approaches to overcome known limitations of state-of-the-art model-based and Markovian approaches in terms of computational complexity and error propagation.

Learning Neural Optimal Interpolation Models and Solvers

no code implementations14 Nov 2022 Maxime Beauchamp, Joseph Thompson, Hugo Georgenthum, Quentin Febvre, Ronan Fablet

The reconstruction of gap-free signals from observation data is a critical challenge for numerous application domains, such as geoscience and space-based earth observation, when the available sensors or the data collection processes lead to irregularly-sampled and noisy observations.

Gaussian Processes

4DVarNet-SSH: end-to-end learning of variational interpolation schemes for nadir and wide-swath satellite altimetry

no code implementations10 Nov 2022 Maxime Beauchamp, Quentin Febvre, Hugo Georgentum, Ronan Fablet

We introduce a parametrization of the 4DVarNet scheme dedicated to the space-time interpolation of satellite altimeter data.

Learning-based estimation of in-situ wind speed from underwater acoustics

no code implementations18 Aug 2022 Matteo Zambra, Dorian Cazau, Nicolas Farrugia, Alexandre Gensse, Sara Pensieri, Roberto Bozzano, Ronan Fablet

As sea surface winds produce sounds that propagate underwater, underwater acoustics recordings can also deliver fine-grained wind-related information.

Retrieval Time Series Analysis

Multimodal 4DVarNets for the reconstruction of sea surface dynamics from SST-SSH synergies

1 code implementation4 Jul 2022 Ronan Fablet, Quentin Febvre, Bertrand Chapron

We introduce a trainable multimodal inversion scheme for the reconstruction of sea surface dynamics from multi-source satellite-derived observations.

A posteriori learning for quasi-geostrophic turbulence parametrization

no code implementations8 Apr 2022 Hugo Frezat, Julien Le Sommer, Ronan Fablet, Guillaume Balarac, Redouane Lguensat

State-of-the-art strategies address the problem as a supervised learning task and optimize algorithms that predict subgrid fluxes based on information from coarse resolution models.

Multimodal learning-based inversion models for the space-time reconstruction of satellite-derived geophysical fields

1 code implementation20 Mar 2022 Ronan Fablet, Bertrand Chapron

For numerous earth observation applications, one may benefit from various satellite sensors to address the reconstruction of some process or information of interest.

Bounded nonlinear forecasts of partially observed geophysical systems with physics-constrained deep learning

no code implementations11 Feb 2022 Said Ouala, Steven L. Brunton, Ananda Pascual, Bertrand Chapron, Fabrice Collard, Lucile Gaultier, Ronan Fablet

The complexity of real-world geophysical systems is often compounded by the fact that the observed measurements depend on hidden variables.

A posteriori learning of quasi-geostrophic turbulence parametrization: an experiment on integration steps

1 code implementation12 Nov 2021 Hugo Frezat, Julien Le Sommer, Ronan Fablet, Guillaume Balarac, Redouane Lguensat

Modeling the subgrid-scale dynamics of reduced models is a long standing open problem that finds application in ocean, atmosphere and climate predictions where direct numerical simulation (DNS) is impossible.

Joint calibration and mapping of satellite altimetry data using trainable variational models

no code implementations7 Oct 2021 Quentin Febvre, Ronan Fablet, Julien Le Sommer, Clément Ubelmann

The proposed framework significantly outperforms the operational state-of-the-art mapping pipeline and truly benefits from wide-swath data to resolve finer scales on the global map as well as in the SWOT sensor geometry.

TrAISformer-A generative transformer for AIS trajectory prediction

1 code implementation8 Sep 2021 Duong Nguyen, Ronan Fablet

The prediction of vessel positions at a specified point in the future is a fundamental aspect of many maritime applications.

Trajectory Forecasting

Learning stochastic dynamical systems with neural networks mimicking the Euler-Maruyama scheme

no code implementations18 May 2021 Noura Dridi, Lucas Drumetz, Ronan Fablet

They are notable for the ability to include a deterministic component of the system and a stochastic one to represent random unknown factors.

Learning Runge-Kutta Integration Schemes for ODE Simulation and Identification

no code implementations11 May 2021 Said Ouala, Laurent Debreu, Ananda Pascual, Bertrand Chapron, Fabrice Collard, Lucile Gaultier, Ronan Fablet

Inevitably, a numerical simulation of a differential equation will then always be distinct from a true analytical solution.

Physical invariance in neural networks for subgrid-scale scalar flux modeling

1 code implementation9 Oct 2020 Hugo Frezat, Guillaume Balarac, Julien Le Sommer, Ronan Fablet, Redouane Lguensat

In this paper we present a new strategy to model the subgrid-scale scalar flux in a three-dimensional turbulent incompressible flow using physics-informed neural networks (NNs).

Variational Deep Learning for the Identification and Reconstruction of Chaotic and Stochastic Dynamical Systems from Noisy and Partial Observations

1 code implementation4 Sep 2020 Duong Nguyen, Said Ouala, Lucas. Drumetz, Ronan Fablet

The data-driven recovery of the unknown governing equations of dynamical systems has recently received an increasing interest.

Detection of Abnormal Vessel Behaviours from AIS data using GeoTrackNet: from the Laboratory to the Ocean

no code implementations12 Aug 2020 Duong Nguyen, Matthieu Simonin, Guillaume Hajduch, Rodolphe Vadaine, Cédric Tedeschi, Ronan Fablet

The constant growth of maritime traffic leads to the need of automatic anomaly detection, which has been attracting great research attention.

Anomaly Detection

Learning Variational Data Assimilation Models and Solvers

2 code implementations25 Jul 2020 Ronan Fablet, Bertrand Chapron, Lucas. Drumetz, Etienne Memin, Olivier Pannekoucke, Francois Rousseau

Intriguingly, we also show that the variational models issued from the true Lorenz-63 and Lorenz-96 ODE representations may not lead to the best reconstruction performance.

Joint learning of variational representations and solvers for inverse problems with partially-observed data

5 code implementations5 Jun 2020 Ronan Fablet, Lucas. Drumetz, Francois Rousseau

The variational cost and the gradient-based solver are both stated as neural networks using automatic differentiation for the latter.

Image Inpainting Time Series +1

Filtering Internal Tides From Wide-Swath Altimeter Data Using Convolutional Neural Networks

1 code implementation3 May 2020 Redouane Lguensat, Ronan Fablet, Julien Le Sommer, Sammy Metref, Emmanuel Cosme, Kaouther Ouenniche, Lucas. Drumetz, Jonathan Gula

The upcoming Surface Water Ocean Topography (SWOT) satellite altimetry mission is expected to yield two-dimensional high-resolution measurements of Sea Surface Height (SSH), thus allowing for a better characterization of the mesoscale and submesoscale eddy field.

PDE-NetGen 1.0: from symbolic PDE representations of physical processes to trainable neural network representations

1 code implementation3 Feb 2020 Olivier Pannekoucke, Ronan Fablet

While deep learning frameworks open avenues in physical science, the design of physically-consistent deep neural network architectures is an open issue.

Learning Generalized Quasi-Geostrophic Models Using Deep Neural Numerical Models

1 code implementation20 Nov 2019 Redouane Lguensat, Julien Le Sommer, Sammy Metref, Emmanuel Cosme, Ronan Fablet

We introduce a new strategy designed to help physicists discover hidden laws governing dynamical systems.

End-to-end learning of energy-based representations for irregularly-sampled signals and images

4 code implementations1 Oct 2019 Ronan Fablet, Lucas. Drumetz, François Rousseau

In this paper, we address the end-to-end learning of representations of signals, images and image sequences from irregularly-sampled data, i. e. when the training data involved missing data.

Time Series Time Series Analysis

Learning Latent Dynamics for Partially-Observed Chaotic Systems

1 code implementation4 Jul 2019 Said Ouala, Duong Nguyen, Lucas. Drumetz, Bertrand Chapron, Ananda Pascual, Fabrice Collard, Lucile Gaultier, Ronan Fablet

This paper addresses the data-driven identification of latent dynamical representations of partially-observed systems, i. e., dynamical systems for which some components are never observed, with an emphasis on forecasting applications, including long-term asymptotic patterns.

EM-like Learning Chaotic Dynamics from Noisy and Partial Observations

no code implementations25 Mar 2019 Duong Nguyen, Said Ouala, Lucas. Drumetz, Ronan Fablet

To solve for the joint inference of the hidden dynamics and of model parameters, we combine neural-network representations and state-of-the-art assimilation schemes.

Recurrent Neural Networks with Stochastic Layers for Acoustic Novelty Detection

1 code implementation13 Feb 2019 Duong Nguyen, Oliver S. Kirsebom, Fábio Frazão, Ronan Fablet, Stan Matwin

In this paper, we adapt Recurrent Neural Networks with Stochastic Layers, which are the state-of-the-art for generating text, music and speech, to the problem of acoustic novelty detection.

Acoustic Novelty Detection Feature Engineering +1

Sea surface temperature prediction and reconstruction using patch-level neural network representations

no code implementations1 Jun 2018 Said Ouala, Cedric Herzet, Ronan Fablet

The forecasting and reconstruction of ocean and atmosphere dynamics from satellite observation time series are key challenges.

Numerical Integration Time Series +1

Residual Networks as Geodesic Flows of Diffeomorphisms

no code implementations24 May 2018 Francois Rousseau, Ronan Fablet

This paper addresses the understanding and characterization of residual networks (ResNet), which are among the state-of-the-art deep learning architectures for a variety of supervised learning problems.

Bilinear residual Neural Network for the identification and forecasting of dynamical systems

no code implementations19 Dec 2017 Ronan Fablet, Said Ouala, Cedric Herzet

Due to the increasing availability of large-scale observation and simulation datasets, data-driven representations arise as efficient and relevant computation representations of dynamical systems for a wide range of applications, where model-driven models based on ordinary differential equation remain the state-of-the-art approaches.

EddyNet: A Deep Neural Network For Pixel-Wise Classification of Oceanic Eddies

1 code implementation10 Nov 2017 Redouane Lguensat, Miao Sun, Ronan Fablet, Evan Mason, Pierre Tandeo, Ge Chen

This work presents EddyNet, a deep learning based architecture for automated eddy detection and classification from Sea Surface Height (SSH) maps provided by the Copernicus Marine and Environment Monitoring Service (CMEMS).

General Classification Oceanic Eddy Classification

Cannot find the paper you are looking for? You can Submit a new open access paper.