Learning Variational Data Assimilation Models and Solvers

07/25/2020
by   Ronan Fablet, et al.
0

This paper addresses variational data assimilation from a learning point of view. Data assimilation aims to reconstruct the time evolution of some state given a series of observations, possibly noisy and irregularly-sampled. Using automatic differentiation tools embedded in deep learning frameworks, we introduce end-to-end neural network architectures for data assimilation. It comprises two key components: a variational model and a gradient-based solver both implemented as neural networks. A key feature of the proposed end-to-end learning architecture is that we may train the NN models using both supervised and unsupervised strategies. Our numerical experiments on Lorenz-63 and Lorenz-96 systems report significant gain w.r.t. a classic gradient-based minimization of the variational cost both in terms of reconstruction performance and optimization complexity. 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. We believe these results may open new research avenues for the specification of assimilation models in geoscience.

READ FULL TEXT
research
06/05/2020

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

Designing appropriate variational regularization schemes is a crucial pa...
research
04/26/2017

Limits of End-to-End Learning

End-to-end learning refers to training a possibly complex learning syste...
research
06/25/2023

A differentiable Gaussian Prototype Layer for explainable Segmentation

We introduce a Gaussian Prototype Layer for gradient-based prototype lea...
research
09/11/2023

Neural Koopman prior for data assimilation

With the increasing availability of large scale datasets, computational ...
research
04/03/2023

Neuroevolution of Recurrent Architectures on Control Tasks

Modern artificial intelligence works typically train the parameters of f...
research
07/01/2019

Model Comparison of Dark Energy models Using Deep Network

This work uses the combination of the variational auto-encoder and the g...
research
07/10/2023

An End-To-End Analysis of Deep Learning-Based Remaining Useful Life Algorithms for Satefy-Critical 5G-Enabled IIoT Networks

Remaining Useful Life (RUL) prediction is a critical task that aims to e...

Please sign up or login with your details

Forgot password? Click here to reset